Created
October 31, 2022 04:37
-
-
Save ezyang/1be7475e7303fb9190bcb7729cbba257 to your computer and use it in GitHub Desktop.
Sweep logs for symbolic-shapes --accuracy --backend inductor --training (TORCHDYNAMO_DYNAMIC_SHAPES=1) - Sun Oct 30 20:55:31 PDT 2022
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Running BERT_pytorch... | |
ERROR:common: | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1041, in check_accuracy | |
new_result = optimized_model_iter_fn(model, example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 945, in run_n_iterations | |
self.model_iter_fn(mod, inputs, collect_outputs=False) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 332, in forward_and_backward_pass | |
cloned_inputs = clone_inputs(inputs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 335, in <graph break in forward_and_backward_pass> | |
pred = mod(*cloned_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "/data/users/ezyang/benchmark/torchbenchmark/models/BERT_pytorch/bert_pytorch/model/bert.py", line 43, in forward | |
x = self.embedding(x, segment_info) | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "/data/users/ezyang/benchmark/torchbenchmark/models/BERT_pytorch/bert_pytorch/model/embedding/bert.py", line 32, in forward | |
x = self.token(sequence) + self.position(sequence) + self.segment(segment_label) | |
File "/data/users/ezyang/benchmark/torchbenchmark/models/BERT_pytorch/bert_pytorch/model/embedding/bert.py", line 32, in <graph break in forward> | |
x = self.token(sequence) + self.position(sequence) + self.segment(segment_label) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 893, in forward | |
return compiled_f( | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 880, in new_func | |
compiled_fn = create_aot_dispatcher_function( | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 600, in create_aot_dispatcher_function | |
aot_dispatch_autograd(flat_fn, fake_flat_tensor_args, aot_config) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 459, in aot_dispatch_autograd | |
compiled_fw_func = aot_config.fw_compiler(fw_module, deduped_flat_args) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 351, in fw_compiler | |
return compile_fx_inner( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/debug_utils.py", line 466, in debug_wrapper | |
compiled_fn = compiler_fn(gm, example_inputs, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/debug.py", line 177, in inner | |
return fn(*args, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/contextlib.py", line 79, in inner | |
return func(*args, **kwds) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 123, in compile_fx_inner | |
compiled_fn = graph.compile_to_fn() | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 346, in compile_to_fn | |
return self.compile_to_module().call | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 332, in compile_to_module | |
code = self.codegen() | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 323, in codegen | |
self.wrapper_code = WrapperCodeGen() | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/codegen/wrapper.py", line 240, in __init__ | |
V.graph.sizevars.codegen(self.prefix, V.graph.graph_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/sizevars.py", line 481, in codegen | |
assert not needed | |
AssertionError | |
TorchDynamo optimized model failed to run because of following error | |
cuda train BERT_pytorch FAIL | |
Running Background_Matting... | |
ERROR:common:RuntimeError: Overloaded torch operator invoked from Python failed to many any schema: | |
aten::arange() expected at most 5 argument(s) but received 7 argument(s). Declaration: aten::arange(Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() expected at most 6 argument(s) but received 7 argument(s). Declaration: aten::arange.start(Scalar start, Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() Expected a value of type 'number' for argument 'end' but instead found type 'Float'. | |
Position: 1 | |
Value: 256.000000000000 | |
Declaration: aten::arange.start_step(Scalar start, Scalar end, Scalar step=1, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
Cast error details: Cannot cast 256.000000000000 to number | |
aten::arange() expected at most 4 argument(s) but received 7 argument(s). Declaration: aten::arange.start_out(Scalar start, Scalar end, Scalar step=1, *, Tensor(a!) out) -> Tensor(a!) | |
aten::arange() expected at most 2 argument(s) but received 7 argument(s). Declaration: aten::arange.out(Scalar end, *, Tensor(a!) out) -> Tensor(a!) | |
target: aten.arange.default | |
args[0]: 256.000000000000 | |
kwargs: {'dtype': torch.float32, 'device': device(type='cuda', index=0), 'pin_memory': False} | |
While executing %arange : [#users=1] = call_function[target=torch.ops.aten.arange.default](args = (%mul_108,), kwargs = {dtype: torch.float32, device: cuda:0, pin_memory: False}) | |
Original traceback: | |
Module stack: {'self_model_al_out': <class 'torch.nn.modules.container.Sequential'>, 'self_model_al_out_0': <class 'torch.nn.modules.upsampling.Upsample'>} | |
File "/data/users/ezyang/benchmark/torchbenchmark/models/Background_Matting/networks.py", line 104, in forward | |
al_out=self.model_al_out(out_dec_al) | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 239, in call_function | |
out = lowerings[target](*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/lowering.py", line 204, in wrapped | |
return decomp_fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/lowering.py", line 1194, in arange | |
return fallback_arange( | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/lowering.py", line 893, in handler | |
result = ir.FallbackKernel.create(kernel, *args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/ir.py", line 2890, in create | |
example_output = kernel( | |
File "/data/users/ezyang/pytorch-tmp/torch/_ops.py", line 446, in __call__ | |
return self._op(*args, **kwargs or {}) | |
RuntimeError: Overloaded torch operator invoked from Python failed to many any schema: | |
aten::arange() expected at most 5 argument(s) but received 7 argument(s). Declaration: aten::arange(Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() expected at most 6 argument(s) but received 7 argument(s). Declaration: aten::arange.start(Scalar start, Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() Expected a value of type 'number' for argument 'end' but instead found type 'Float'. | |
Position: 1 | |
Value: 256.000000000000 | |
Declaration: aten::arange.start_step(Scalar start, Scalar end, Scalar step=1, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
Cast error details: Cannot cast 256.000000000000 to number | |
aten::arange() expected at most 4 argument(s) but received 7 argument(s). Declaration: aten::arange.start_out(Scalar start, Scalar end, Scalar step=1, *, Tensor(a!) out) -> Tensor(a!) | |
aten::arange() expected at most 2 argument(s) but received 7 argument(s). Declaration: aten::arange.out(Scalar end, *, Tensor(a!) out) -> Tensor(a!) | |
The above exception was the direct cause of the following exception: | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1041, in check_accuracy | |
new_result = optimized_model_iter_fn(model, example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 945, in run_n_iterations | |
self.model_iter_fn(mod, inputs, collect_outputs=False) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 332, in forward_and_backward_pass | |
cloned_inputs = clone_inputs(inputs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 335, in <graph break in forward_and_backward_pass> | |
pred = mod(*cloned_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "/data/users/ezyang/benchmark/torchbenchmark/models/Background_Matting/networks.py", line 91, in forward | |
def forward(self, image,back,seg,multi): | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 893, in forward | |
return compiled_f( | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 880, in new_func | |
compiled_fn = create_aot_dispatcher_function( | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 600, in create_aot_dispatcher_function | |
aot_dispatch_autograd(flat_fn, fake_flat_tensor_args, aot_config) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 459, in aot_dispatch_autograd | |
compiled_fw_func = aot_config.fw_compiler(fw_module, deduped_flat_args) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 351, in fw_compiler | |
return compile_fx_inner( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/debug_utils.py", line 466, in debug_wrapper | |
compiled_fn = compiler_fn(gm, example_inputs, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/debug.py", line 177, in inner | |
return fn(*args, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/contextlib.py", line 79, in inner | |
return func(*args, **kwds) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 122, in compile_fx_inner | |
graph.run(*example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 129, in run | |
return super().run(*args) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 130, in run | |
self.env[node] = self.run_node(node) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 299, in run_node | |
result = super().run_node(n) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 171, in run_node | |
return getattr(self, n.op)(n.target, args, kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 242, in call_function | |
raise LoweringException(e, target, args, kwargs) from e | |
torch._inductor.exc.LoweringException: RuntimeError: Overloaded torch operator invoked from Python failed to many any schema: | |
aten::arange() expected at most 5 argument(s) but received 7 argument(s). Declaration: aten::arange(Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() expected at most 6 argument(s) but received 7 argument(s). Declaration: aten::arange.start(Scalar start, Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() Expected a value of type 'number' for argument 'end' but instead found type 'Float'. | |
Position: 1 | |
Value: 256.000000000000 | |
Declaration: aten::arange.start_step(Scalar start, Scalar end, Scalar step=1, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
Cast error details: Cannot cast 256.000000000000 to number | |
aten::arange() expected at most 4 argument(s) but received 7 argument(s). Declaration: aten::arange.start_out(Scalar start, Scalar end, Scalar step=1, *, Tensor(a!) out) -> Tensor(a!) | |
aten::arange() expected at most 2 argument(s) but received 7 argument(s). Declaration: aten::arange.out(Scalar end, *, Tensor(a!) out) -> Tensor(a!) | |
target: aten.arange.default | |
args[0]: 256.000000000000 | |
kwargs: {'dtype': torch.float32, 'device': device(type='cuda', index=0), 'pin_memory': False} | |
While executing %arange : [#users=1] = call_function[target=torch.ops.aten.arange.default](args = (%mul_108,), kwargs = {dtype: torch.float32, device: cuda:0, pin_memory: False}) | |
Original traceback: | |
Module stack: {'self_model_al_out': <class 'torch.nn.modules.container.Sequential'>, 'self_model_al_out_0': <class 'torch.nn.modules.upsampling.Upsample'>} | |
File "/data/users/ezyang/benchmark/torchbenchmark/models/Background_Matting/networks.py", line 104, in forward | |
al_out=self.model_al_out(out_dec_al) | |
TorchDynamo optimized model failed to run because of following error | |
cuda train Background_Matting FAIL | |
WARNING:root:DALLE2_pytorch failed to load | |
Eager model failed to run | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 903, in validate_model | |
self.model_iter_fn(model, example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 337, in forward_and_backward_pass | |
self.grad_scaler.scale(loss).backward() | |
File "/data/users/ezyang/pytorch-tmp/torch/_tensor.py", line 488, in backward | |
torch.autograd.backward( | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/__init__.py", line 197, in backward | |
Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass | |
RuntimeError: element 0 of tensors does not require grad and does not have a grad_fn | |
During handling of the above exception, another exception occurred: | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1746, in main | |
device, name, model, example_inputs, batch_size = runner.load_model( | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 282, in load_model | |
self.validate_model(model, example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 905, in validate_model | |
raise NotImplementedError("Eager model failed to run") | |
NotImplementedError: Eager model failed to run | |
Running LearningToPaint... | |
ERROR:common:'int' object has no attribute 'size' | |
While executing %sym_size : [#users=2] = placeholder[target=sym_size] | |
Original traceback: | |
Module stack: {} | |
File "/data/users/ezyang/benchmark/torchbenchmark/models/LearningToPaint/baseline/DRL/actor.py", line 111, in forward | |
x = x.view(x.size(0), -1) | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1041, in check_accuracy | |
new_result = optimized_model_iter_fn(model, example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 945, in run_n_iterations | |
self.model_iter_fn(mod, inputs, collect_outputs=False) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 332, in forward_and_backward_pass | |
cloned_inputs = clone_inputs(inputs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 337, in <graph break in forward_and_backward_pass> | |
self.grad_scaler.scale(loss).backward() | |
File "/data/users/ezyang/pytorch-tmp/torch/_tensor.py", line 488, in backward | |
torch.autograd.backward( | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/__init__.py", line 197, in backward | |
Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/function.py", line 270, in apply | |
return user_fn(self, *args) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 499, in backward | |
CompiledFunction.compiled_bw = aot_config.bw_compiler( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/optimizations/backends.py", line 555, in _wrapped_bw_compiler | |
return disable(bw_compiler(*args, **kwargs)) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 362, in bw_compiler | |
return compile_fx_inner( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/debug_utils.py", line 466, in debug_wrapper | |
compiled_fn = compiler_fn(gm, example_inputs, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/debug.py", line 177, in inner | |
return fn(*args, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/contextlib.py", line 79, in inner | |
return func(*args, **kwds) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 122, in compile_fx_inner | |
graph.run(*example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 129, in run | |
return super().run(*args) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 130, in run | |
self.env[node] = self.run_node(node) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 299, in run_node | |
result = super().run_node(n) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 171, in run_node | |
return getattr(self, n.op)(n.target, args, kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 199, in placeholder | |
sizes, strides = self.static_sizes_strides(example) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 53, in static_sizes_strides | |
size = [sympy.Integer(i) for i in ex.size()] | |
AttributeError: 'int' object has no attribute 'size' | |
While executing %sym_size : [#users=2] = placeholder[target=sym_size] | |
Original traceback: | |
Module stack: {} | |
File "/data/users/ezyang/benchmark/torchbenchmark/models/LearningToPaint/baseline/DRL/actor.py", line 111, in forward | |
x = x.view(x.size(0), -1) | |
TorchDynamo optimized model failed to run because of following error | |
cuda train LearningToPaint FAIL | |
Running Super_SloMo... | |
ERROR:common:RuntimeError: Overloaded torch operator invoked from Python failed to many any schema: | |
aten::arange() expected at most 5 argument(s) but received 7 argument(s). Declaration: aten::arange(Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() expected at most 6 argument(s) but received 7 argument(s). Declaration: aten::arange.start(Scalar start, Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() Expected a value of type 'number' for argument 'end' but instead found type 'Float'. | |
Position: 1 | |
Value: 22.0000000000000 | |
Declaration: aten::arange.start_step(Scalar start, Scalar end, Scalar step=1, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
Cast error details: Cannot cast 22.0000000000000 to number | |
aten::arange() expected at most 4 argument(s) but received 7 argument(s). Declaration: aten::arange.start_out(Scalar start, Scalar end, Scalar step=1, *, Tensor(a!) out) -> Tensor(a!) | |
aten::arange() expected at most 2 argument(s) but received 7 argument(s). Declaration: aten::arange.out(Scalar end, *, Tensor(a!) out) -> Tensor(a!) | |
target: aten.arange.default | |
args[0]: 22.0000000000000 | |
kwargs: {'dtype': torch.float32, 'device': device(type='cuda', index=0), 'pin_memory': False} | |
While executing %arange : [#users=1] = call_function[target=torch.ops.aten.arange.default](args = (%mul_12,), kwargs = {dtype: torch.float32, device: cuda:0, pin_memory: False}) | |
Original traceback: | |
Module stack: {'self_flowComp': <class 'torchbenchmark.models.Super_SloMo.slomo_model.UNet'>, 'self_flowComp_up1': <class 'torchbenchmark.models.Super_SloMo.slomo_model.up'>} | |
File "/data/users/ezyang/benchmark/torchbenchmark/models/Super_SloMo/slomo_model.py", line 130, in forward | |
x = F.interpolate(x, scale_factor=2., mode='bilinear') | |
| File "/data/users/ezyang/benchmark/torchbenchmark/models/Super_SloMo/slomo_model.py", line 204, in forward | |
x = self.up1(x, s5) | |
| File "/data/users/ezyang/benchmark/torchbenchmark/models/Super_SloMo/model_wrapper.py", line 28, in forward | |
flowOut = self.flowComp(torch.cat((I0, I1), dim=1)) | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 239, in call_function | |
out = lowerings[target](*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/lowering.py", line 204, in wrapped | |
return decomp_fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/lowering.py", line 1194, in arange | |
return fallback_arange( | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/lowering.py", line 893, in handler | |
result = ir.FallbackKernel.create(kernel, *args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/ir.py", line 2890, in create | |
example_output = kernel( | |
File "/data/users/ezyang/pytorch-tmp/torch/_ops.py", line 446, in __call__ | |
return self._op(*args, **kwargs or {}) | |
RuntimeError: Overloaded torch operator invoked from Python failed to many any schema: | |
aten::arange() expected at most 5 argument(s) but received 7 argument(s). Declaration: aten::arange(Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() expected at most 6 argument(s) but received 7 argument(s). Declaration: aten::arange.start(Scalar start, Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() Expected a value of type 'number' for argument 'end' but instead found type 'Float'. | |
Position: 1 | |
Value: 22.0000000000000 | |
Declaration: aten::arange.start_step(Scalar start, Scalar end, Scalar step=1, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
Cast error details: Cannot cast 22.0000000000000 to number | |
aten::arange() expected at most 4 argument(s) but received 7 argument(s). Declaration: aten::arange.start_out(Scalar start, Scalar end, Scalar step=1, *, Tensor(a!) out) -> Tensor(a!) | |
aten::arange() expected at most 2 argument(s) but received 7 argument(s). Declaration: aten::arange.out(Scalar end, *, Tensor(a!) out) -> Tensor(a!) | |
The above exception was the direct cause of the following exception: | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1041, in check_accuracy | |
new_result = optimized_model_iter_fn(model, example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 945, in run_n_iterations | |
self.model_iter_fn(mod, inputs, collect_outputs=False) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 332, in forward_and_backward_pass | |
cloned_inputs = clone_inputs(inputs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 335, in <graph break in forward_and_backward_pass> | |
pred = mod(*cloned_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "/data/users/ezyang/benchmark/torchbenchmark/models/Super_SloMo/model_wrapper.py", line 26, in forward | |
def forward(self, trainFrameIndex, I0, I1, IFrame): | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 893, in forward | |
return compiled_f( | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 880, in new_func | |
compiled_fn = create_aot_dispatcher_function( | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 600, in create_aot_dispatcher_function | |
aot_dispatch_autograd(flat_fn, fake_flat_tensor_args, aot_config) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 459, in aot_dispatch_autograd | |
compiled_fw_func = aot_config.fw_compiler(fw_module, deduped_flat_args) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 351, in fw_compiler | |
return compile_fx_inner( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/debug_utils.py", line 466, in debug_wrapper | |
compiled_fn = compiler_fn(gm, example_inputs, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/debug.py", line 177, in inner | |
return fn(*args, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/contextlib.py", line 79, in inner | |
return func(*args, **kwds) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 122, in compile_fx_inner | |
graph.run(*example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 129, in run | |
return super().run(*args) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 130, in run | |
self.env[node] = self.run_node(node) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 299, in run_node | |
result = super().run_node(n) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 171, in run_node | |
return getattr(self, n.op)(n.target, args, kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 242, in call_function | |
raise LoweringException(e, target, args, kwargs) from e | |
torch._inductor.exc.LoweringException: RuntimeError: Overloaded torch operator invoked from Python failed to many any schema: | |
aten::arange() expected at most 5 argument(s) but received 7 argument(s). Declaration: aten::arange(Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() expected at most 6 argument(s) but received 7 argument(s). Declaration: aten::arange.start(Scalar start, Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() Expected a value of type 'number' for argument 'end' but instead found type 'Float'. | |
Position: 1 | |
Value: 22.0000000000000 | |
Declaration: aten::arange.start_step(Scalar start, Scalar end, Scalar step=1, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
Cast error details: Cannot cast 22.0000000000000 to number | |
aten::arange() expected at most 4 argument(s) but received 7 argument(s). Declaration: aten::arange.start_out(Scalar start, Scalar end, Scalar step=1, *, Tensor(a!) out) -> Tensor(a!) | |
aten::arange() expected at most 2 argument(s) but received 7 argument(s). Declaration: aten::arange.out(Scalar end, *, Tensor(a!) out) -> Tensor(a!) | |
target: aten.arange.default | |
args[0]: 22.0000000000000 | |
kwargs: {'dtype': torch.float32, 'device': device(type='cuda', index=0), 'pin_memory': False} | |
While executing %arange : [#users=1] = call_function[target=torch.ops.aten.arange.default](args = (%mul_12,), kwargs = {dtype: torch.float32, device: cuda:0, pin_memory: False}) | |
Original traceback: | |
Module stack: {'self_flowComp': <class 'torchbenchmark.models.Super_SloMo.slomo_model.UNet'>, 'self_flowComp_up1': <class 'torchbenchmark.models.Super_SloMo.slomo_model.up'>} | |
File "/data/users/ezyang/benchmark/torchbenchmark/models/Super_SloMo/slomo_model.py", line 130, in forward | |
x = F.interpolate(x, scale_factor=2., mode='bilinear') | |
| File "/data/users/ezyang/benchmark/torchbenchmark/models/Super_SloMo/slomo_model.py", line 204, in forward | |
x = self.up1(x, s5) | |
| File "/data/users/ezyang/benchmark/torchbenchmark/models/Super_SloMo/model_wrapper.py", line 28, in forward | |
flowOut = self.flowComp(torch.cat((I0, I1), dim=1)) | |
TorchDynamo optimized model failed to run because of following error | |
cuda train Super_SloMo FAIL | |
Running alexnet... | |
ERROR:common:'int' object has no attribute 'size' | |
While executing %sym_size : [#users=1] = placeholder[target=sym_size] | |
Original traceback: | |
Module stack: {} | |
File "/data/users/ezyang/vision/torchvision/models/alexnet.py", line 50, in forward | |
x = torch.flatten(x, 1) | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1041, in check_accuracy | |
new_result = optimized_model_iter_fn(model, example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 945, in run_n_iterations | |
self.model_iter_fn(mod, inputs, collect_outputs=False) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 332, in forward_and_backward_pass | |
cloned_inputs = clone_inputs(inputs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 337, in <graph break in forward_and_backward_pass> | |
self.grad_scaler.scale(loss).backward() | |
File "/data/users/ezyang/pytorch-tmp/torch/_tensor.py", line 488, in backward | |
torch.autograd.backward( | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/__init__.py", line 197, in backward | |
Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/function.py", line 270, in apply | |
return user_fn(self, *args) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 499, in backward | |
CompiledFunction.compiled_bw = aot_config.bw_compiler( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/optimizations/backends.py", line 555, in _wrapped_bw_compiler | |
return disable(bw_compiler(*args, **kwargs)) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 362, in bw_compiler | |
return compile_fx_inner( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/debug_utils.py", line 466, in debug_wrapper | |
compiled_fn = compiler_fn(gm, example_inputs, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/debug.py", line 177, in inner | |
return fn(*args, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/contextlib.py", line 79, in inner | |
return func(*args, **kwds) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 122, in compile_fx_inner | |
graph.run(*example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 129, in run | |
return super().run(*args) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 130, in run | |
self.env[node] = self.run_node(node) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 299, in run_node | |
result = super().run_node(n) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 171, in run_node | |
return getattr(self, n.op)(n.target, args, kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 199, in placeholder | |
sizes, strides = self.static_sizes_strides(example) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 53, in static_sizes_strides | |
size = [sympy.Integer(i) for i in ex.size()] | |
AttributeError: 'int' object has no attribute 'size' | |
While executing %sym_size : [#users=1] = placeholder[target=sym_size] | |
Original traceback: | |
Module stack: {} | |
File "/data/users/ezyang/vision/torchvision/models/alexnet.py", line 50, in forward | |
x = torch.flatten(x, 1) | |
TorchDynamo optimized model failed to run because of following error | |
cuda train alexnet FAIL | |
Running attention_is_all_you_need_pytorch... | |
ERROR:common:'int' object has no attribute 'size' | |
While executing %sym_size : [#users=2] = placeholder[target=sym_size] | |
Original traceback: | |
Module stack: {'self_trg_word_prj': <class 'torch.nn.modules.linear.Linear'>} | |
File "/data/users/ezyang/benchmark/torchbenchmark/models/attention_is_all_you_need_pytorch/transformer/Models.py", line 173, in <graph break in forward> | |
seq_logit = self.trg_word_prj(dec_output) * self.x_logit_scale | |
Gradient addition node due to multiple use of tensor around: | |
Module stack: {'self_src_word_emb': <class 'torch.nn.modules.sparse.Embedding'>} | |
File "/data/users/ezyang/benchmark/torchbenchmark/models/attention_is_all_you_need_pytorch/transformer/Models.py", line 71, in forward | |
enc_output = self.dropout(self.position_enc(self.src_word_emb(src_seq))) | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1041, in check_accuracy | |
new_result = optimized_model_iter_fn(model, example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 945, in run_n_iterations | |
self.model_iter_fn(mod, inputs, collect_outputs=False) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 332, in forward_and_backward_pass | |
cloned_inputs = clone_inputs(inputs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 337, in <graph break in forward_and_backward_pass> | |
self.grad_scaler.scale(loss).backward() | |
File "/data/users/ezyang/pytorch-tmp/torch/_tensor.py", line 488, in backward | |
torch.autograd.backward( | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/__init__.py", line 197, in backward | |
Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/function.py", line 270, in apply | |
return user_fn(self, *args) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 499, in backward | |
CompiledFunction.compiled_bw = aot_config.bw_compiler( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/optimizations/backends.py", line 555, in _wrapped_bw_compiler | |
return disable(bw_compiler(*args, **kwargs)) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 362, in bw_compiler | |
return compile_fx_inner( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/debug_utils.py", line 466, in debug_wrapper | |
compiled_fn = compiler_fn(gm, example_inputs, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/debug.py", line 177, in inner | |
return fn(*args, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/contextlib.py", line 79, in inner | |
return func(*args, **kwds) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 122, in compile_fx_inner | |
graph.run(*example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 129, in run | |
return super().run(*args) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 130, in run | |
self.env[node] = self.run_node(node) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 299, in run_node | |
result = super().run_node(n) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 171, in run_node | |
return getattr(self, n.op)(n.target, args, kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 199, in placeholder | |
sizes, strides = self.static_sizes_strides(example) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 53, in static_sizes_strides | |
size = [sympy.Integer(i) for i in ex.size()] | |
AttributeError: 'int' object has no attribute 'size' | |
While executing %sym_size : [#users=2] = placeholder[target=sym_size] | |
Original traceback: | |
Module stack: {'self_trg_word_prj': <class 'torch.nn.modules.linear.Linear'>} | |
File "/data/users/ezyang/benchmark/torchbenchmark/models/attention_is_all_you_need_pytorch/transformer/Models.py", line 173, in <graph break in forward> | |
seq_logit = self.trg_word_prj(dec_output) * self.x_logit_scale | |
Gradient addition node due to multiple use of tensor around: | |
Module stack: {'self_src_word_emb': <class 'torch.nn.modules.sparse.Embedding'>} | |
File "/data/users/ezyang/benchmark/torchbenchmark/models/attention_is_all_you_need_pytorch/transformer/Models.py", line 71, in forward | |
enc_output = self.dropout(self.position_enc(self.src_word_emb(src_seq))) | |
TorchDynamo optimized model failed to run because of following error | |
cuda train attention_is_all_you_need_pytorch FAIL | |
Running dcgan... | |
cuda train dcgan PASS | |
Running densenet121... | |
ERROR:common:'int' object has no attribute 'size' | |
While executing %sym_size : [#users=2] = placeholder[target=sym_size] | |
Original traceback: | |
Module stack: {} | |
File "/data/users/ezyang/vision/torchvision/models/densenet.py", line 217, in forward | |
out = torch.flatten(out, 1) | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1041, in check_accuracy | |
new_result = optimized_model_iter_fn(model, example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 945, in run_n_iterations | |
self.model_iter_fn(mod, inputs, collect_outputs=False) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 332, in forward_and_backward_pass | |
cloned_inputs = clone_inputs(inputs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 337, in <graph break in forward_and_backward_pass> | |
self.grad_scaler.scale(loss).backward() | |
File "/data/users/ezyang/pytorch-tmp/torch/_tensor.py", line 488, in backward | |
torch.autograd.backward( | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/__init__.py", line 197, in backward | |
Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/function.py", line 270, in apply | |
return user_fn(self, *args) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 499, in backward | |
CompiledFunction.compiled_bw = aot_config.bw_compiler( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/optimizations/backends.py", line 555, in _wrapped_bw_compiler | |
return disable(bw_compiler(*args, **kwargs)) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 362, in bw_compiler | |
return compile_fx_inner( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/debug_utils.py", line 466, in debug_wrapper | |
compiled_fn = compiler_fn(gm, example_inputs, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/debug.py", line 177, in inner | |
return fn(*args, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/contextlib.py", line 79, in inner | |
return func(*args, **kwds) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 122, in compile_fx_inner | |
graph.run(*example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 129, in run | |
return super().run(*args) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 130, in run | |
self.env[node] = self.run_node(node) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 299, in run_node | |
result = super().run_node(n) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 171, in run_node | |
return getattr(self, n.op)(n.target, args, kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 199, in placeholder | |
sizes, strides = self.static_sizes_strides(example) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 53, in static_sizes_strides | |
size = [sympy.Integer(i) for i in ex.size()] | |
AttributeError: 'int' object has no attribute 'size' | |
While executing %sym_size : [#users=2] = placeholder[target=sym_size] | |
Original traceback: | |
Module stack: {} | |
File "/data/users/ezyang/vision/torchvision/models/densenet.py", line 217, in forward | |
out = torch.flatten(out, 1) | |
TorchDynamo optimized model failed to run because of following error | |
cuda train densenet121 FAIL | |
WARNING:root:detectron2_fcos_r_50_fpn failed to load | |
FCOS train is not supported by upstream detectron2. See GH Issue: https://github.com/facebookresearch/detectron2/issues/4369. | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1746, in main | |
device, name, model, example_inputs, batch_size = runner.load_model( | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 251, in load_model | |
benchmark = benchmark_cls( | |
File "/data/users/ezyang/benchmark/torchbenchmark/util/model.py", line 18, in __call__ | |
obj = type.__call__(cls, *args, **kwargs) | |
File "/data/users/ezyang/benchmark/torchbenchmark/models/detectron2_fcos_r_50_fpn/__init__.py", line 15, in __init__ | |
super().__init__(variant="COCO-Detection/fcos_R_50_FPN_1x.py", test=test, device=device, | |
File "/data/users/ezyang/benchmark/torchbenchmark/util/framework/detectron2/model_factory.py", line 100, in __init__ | |
loader = self.setup_train(cfg, args) | |
File "/data/users/ezyang/benchmark/torchbenchmark/util/framework/detectron2/model_factory.py", line 110, in setup_train | |
raise NotImplementedError("FCOS train is not supported by upstream detectron2. " \ | |
NotImplementedError: FCOS train is not supported by upstream detectron2. See GH Issue: https://github.com/facebookresearch/detectron2/issues/4369. | |
WARNING:root:detectron2_maskrcnn_r_50_c4 failed to load | |
Eager model failed to run | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 903, in validate_model | |
self.model_iter_fn(model, example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 336, in forward_and_backward_pass | |
loss = self.compute_loss(pred) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 326, in compute_loss | |
return reduce_to_scalar_loss(pred) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/testing.py", line 87, in reduce_to_scalar_loss | |
return sum([reduce_to_scalar_loss(x) for x in out]) / len(out) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/testing.py", line 87, in <listcomp> | |
return sum([reduce_to_scalar_loss(x) for x in out]) / len(out) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/testing.py", line 97, in reduce_to_scalar_loss | |
return sum([reduce_to_scalar_loss(value) for value in out.values()]) / len( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/testing.py", line 97, in <listcomp> | |
return sum([reduce_to_scalar_loss(value) for value in out.values()]) / len( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/testing.py", line 102, in reduce_to_scalar_loss | |
raise NotImplementedError("Don't know how to reduce", type(out)) | |
NotImplementedError: ("Don't know how to reduce", <class 'detectron2.structures.instances.Instances'>) | |
During handling of the above exception, another exception occurred: | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1746, in main | |
device, name, model, example_inputs, batch_size = runner.load_model( | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 282, in load_model | |
self.validate_model(model, example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 905, in validate_model | |
raise NotImplementedError("Eager model failed to run") | |
NotImplementedError: Eager model failed to run | |
Running torchbench.py dlrm... | |
ERROR:common:Failed running <class 'range'>(*(9,), **{}): | |
'SymInt' object cannot be interpreted as an integer | |
(scroll up for backtrace) | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/variables/tensor.py", line 53, in _run_node | |
return node.target(*args, **kwargs) | |
TypeError: 'SymInt' object cannot be interpreted as an integer | |
The above exception was the direct cause of the following exception: | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1041, in check_accuracy | |
new_result = optimized_model_iter_fn(model, example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 945, in run_n_iterations | |
self.model_iter_fn(mod, inputs, collect_outputs=False) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 332, in forward_and_backward_pass | |
cloned_inputs = clone_inputs(inputs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 335, in <graph break in forward_and_backward_pass> | |
pred = mod(*cloned_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 240, in catch_errors | |
return callback(frame, cache_size) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/convert_frame.py", line 437, in _convert_frame | |
result = inner_convert(frame, cache_size) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/convert_frame.py", line 112, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/convert_frame.py", line 319, in _convert_frame_assert | |
return _compile( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/convert_frame.py", line 374, in _compile | |
out_code = transform_code_object(code, transform) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/bytecode_transformation.py", line 341, in transform_code_object | |
transformations(instructions, code_options) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/convert_frame.py", line 362, in transform | |
tracer.run() | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/symbolic_convert.py", line 1466, in run | |
super().run() | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/symbolic_convert.py", line 352, in run | |
and self.step() | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/symbolic_convert.py", line 325, in step | |
getattr(self, inst.opname)(inst) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/symbolic_convert.py", line 177, in wrapper | |
return inner_fn(self, inst) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/symbolic_convert.py", line 770, in CALL_FUNCTION | |
self.call_function(fn, args, {}) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/symbolic_convert.py", line 267, in call_function | |
self.push(fn.call_function(self, args, kwargs)) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/variables/functions.py", line 221, in call_function | |
return super().call_function(tx, args, kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/variables/functions.py", line 191, in call_function | |
return super(UserFunctionVariable, self).call_function(tx, args, kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/variables/functions.py", line 62, in call_function | |
return tx.inline_user_function_return( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/symbolic_convert.py", line 296, in inline_user_function_return | |
result = InliningInstructionTranslator.inline_call(self, fn, args, kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/symbolic_convert.py", line 1538, in inline_call | |
return cls.inline_call_(parent, func, args, kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/symbolic_convert.py", line 1592, in inline_call_ | |
tracer.run() | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/symbolic_convert.py", line 352, in run | |
and self.step() | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/symbolic_convert.py", line 325, in step | |
getattr(self, inst.opname)(inst) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/symbolic_convert.py", line 177, in wrapper | |
return inner_fn(self, inst) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/symbolic_convert.py", line 770, in CALL_FUNCTION | |
self.call_function(fn, args, {}) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/symbolic_convert.py", line 267, in call_function | |
self.push(fn.call_function(self, args, kwargs)) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/variables/functions.py", line 221, in call_function | |
return super().call_function(tx, args, kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/variables/functions.py", line 191, in call_function | |
return super(UserFunctionVariable, self).call_function(tx, args, kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/variables/functions.py", line 62, in call_function | |
return tx.inline_user_function_return( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/symbolic_convert.py", line 296, in inline_user_function_return | |
result = InliningInstructionTranslator.inline_call(self, fn, args, kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/symbolic_convert.py", line 1538, in inline_call | |
return cls.inline_call_(parent, func, args, kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/symbolic_convert.py", line 1592, in inline_call_ | |
tracer.run() | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/symbolic_convert.py", line 352, in run | |
and self.step() | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/symbolic_convert.py", line 325, in step | |
getattr(self, inst.opname)(inst) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/symbolic_convert.py", line 177, in wrapper | |
return inner_fn(self, inst) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/symbolic_convert.py", line 770, in CALL_FUNCTION | |
self.call_function(fn, args, {}) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/symbolic_convert.py", line 267, in call_function | |
self.push(fn.call_function(self, args, kwargs)) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/variables/builtin.py", line 369, in call_function | |
return DynamicShapeVariable.create(tx, proxy, None, **options) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/variables/tensor.py", line 634, in create | |
dyn_shape = _get_fake_value(proxy.node, tx) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/variables/tensor.py", line 131, in _get_fake_value | |
return wrap_fake_exception( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 709, in wrap_fake_exception | |
return fn() | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/variables/tensor.py", line 132, in <lambda> | |
lambda: _run_node(tx.output, node, args, kwargs, nnmodule) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/variables/tensor.py", line 62, in _run_node | |
raise RuntimeError( | |
RuntimeError: Failed running <class 'range'>(*(9,), **{}): | |
'SymInt' object cannot be interpreted as an integer | |
(scroll up for backtrace) | |
TorchDynamo optimized model failed to run because of following error | |
cuda train dlrm FAIL | |
/data/users/ezyang/pytorch-tmp/torch/utils/tensorboard/__init__.py:4: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead. | |
if not hasattr(tensorboard, "__version__") or LooseVersion( | |
/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/gym/core.py:317: DeprecationWarning: [33mWARN: Initializing wrapper in old step API which returns one bool instead of two. It is recommended to set `new_step_api=True` to use new step API. This will be the default behaviour in future.[0m | |
deprecation( | |
Running torchbench.py drq... | |
cuda train drq FAIL (TIMEOUT) | |
Running torchbench.py fastNLP_Bert... | |
[2022-10-30 21:11:19,106] torch._inductor.ir: [WARNING] DeviceCopy | |
ERROR:common:RuntimeError: Overloaded torch operator invoked from Python failed to many any schema: | |
aten::arange() expected at most 5 argument(s) but received 7 argument(s). Declaration: aten::arange(Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() expected at most 6 argument(s) but received 7 argument(s). Declaration: aten::arange.start(Scalar start, Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() Expected a value of type 'number' for argument 'end' but instead found type 'Symbol'. | |
Position: 1 | |
Value: s1 | |
Declaration: aten::arange.start_step(Scalar start, Scalar end, Scalar step=1, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
Cast error details: Cannot cast s1 to number | |
aten::arange() expected at most 4 argument(s) but received 7 argument(s). Declaration: aten::arange.start_out(Scalar start, Scalar end, Scalar step=1, *, Tensor(a!) out) -> Tensor(a!) | |
aten::arange() expected at most 2 argument(s) but received 7 argument(s). Declaration: aten::arange.out(Scalar end, *, Tensor(a!) out) -> Tensor(a!) | |
target: aten.arange.default | |
args[0]: s1 | |
kwargs: {'dtype': torch.int64, 'device': device(type='cuda', index=0), 'pin_memory': False} | |
While executing %arange : [#users=1] = call_function[target=torch.ops.aten.arange.default](args = (%sym_size,), kwargs = {dtype: torch.int64, device: cuda:0, pin_memory: False}) | |
Original traceback: | |
Module stack: {'self_embeddings': <class 'fastNLP.modules.encoder.bert.BertEmbeddings'>} | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/fastNLP/modules/encoder/bert.py", line 230, in forward | |
position_ids = torch.arange(seq_length, dtype=torch.long, device=input_ids.device) | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/fastNLP/modules/encoder/bert.py", line 512, in forward | |
embedding_output = self.embeddings(input_ids, token_type_ids) | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 239, in call_function | |
out = lowerings[target](*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/lowering.py", line 204, in wrapped | |
return decomp_fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/lowering.py", line 1194, in arange | |
return fallback_arange( | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/lowering.py", line 893, in handler | |
result = ir.FallbackKernel.create(kernel, *args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/ir.py", line 2890, in create | |
example_output = kernel( | |
File "/data/users/ezyang/pytorch-tmp/torch/_ops.py", line 446, in __call__ | |
return self._op(*args, **kwargs or {}) | |
RuntimeError: Overloaded torch operator invoked from Python failed to many any schema: | |
aten::arange() expected at most 5 argument(s) but received 7 argument(s). Declaration: aten::arange(Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() expected at most 6 argument(s) but received 7 argument(s). Declaration: aten::arange.start(Scalar start, Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() Expected a value of type 'number' for argument 'end' but instead found type 'Symbol'. | |
Position: 1 | |
Value: s1 | |
Declaration: aten::arange.start_step(Scalar start, Scalar end, Scalar step=1, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
Cast error details: Cannot cast s1 to number | |
aten::arange() expected at most 4 argument(s) but received 7 argument(s). Declaration: aten::arange.start_out(Scalar start, Scalar end, Scalar step=1, *, Tensor(a!) out) -> Tensor(a!) | |
aten::arange() expected at most 2 argument(s) but received 7 argument(s). Declaration: aten::arange.out(Scalar end, *, Tensor(a!) out) -> Tensor(a!) | |
The above exception was the direct cause of the following exception: | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1041, in check_accuracy | |
new_result = optimized_model_iter_fn(model, example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 945, in run_n_iterations | |
self.model_iter_fn(mod, inputs, collect_outputs=False) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 332, in forward_and_backward_pass | |
cloned_inputs = clone_inputs(inputs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 335, in <graph break in forward_and_backward_pass> | |
pred = mod(*cloned_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/fastNLP/models/bert.py", line 265, in forward | |
sequence_output = self.bert(words) | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/fastNLP/embeddings/bert_embedding.py", line 137, in forward | |
outputs = self.model(words) | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/fastNLP/embeddings/bert_embedding.py", line 445, in forward | |
max_word_piece_length = batch_word_pieces_length.sum(dim=-1).max().item() # 表示word piece的长度(包括padding) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/fastNLP/embeddings/bert_embedding.py", line 462, in <graph break in forward> | |
word_indexes = words.cpu().numpy() | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/fastNLP/embeddings/bert_embedding.py", line 482, in <graph break in forward> | |
bert_outputs, pooled_cls = self.encoder(word_pieces, token_type_ids=token_type_ids, attention_mask=attn_masks, | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/fastNLP/modules/encoder/bert.py", line 480, in forward | |
def forward(self, input_ids, token_type_ids=None, attention_mask=None, output_all_encoded_layers=True): | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 893, in forward | |
return compiled_f( | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 880, in new_func | |
compiled_fn = create_aot_dispatcher_function( | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 600, in create_aot_dispatcher_function | |
aot_dispatch_autograd(flat_fn, fake_flat_tensor_args, aot_config) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 459, in aot_dispatch_autograd | |
compiled_fw_func = aot_config.fw_compiler(fw_module, deduped_flat_args) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 351, in fw_compiler | |
return compile_fx_inner( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/debug_utils.py", line 466, in debug_wrapper | |
compiled_fn = compiler_fn(gm, example_inputs, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/debug.py", line 177, in inner | |
return fn(*args, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/contextlib.py", line 79, in inner | |
return func(*args, **kwds) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 122, in compile_fx_inner | |
graph.run(*example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 129, in run | |
return super().run(*args) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 130, in run | |
self.env[node] = self.run_node(node) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 299, in run_node | |
result = super().run_node(n) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 171, in run_node | |
return getattr(self, n.op)(n.target, args, kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 242, in call_function | |
raise LoweringException(e, target, args, kwargs) from e | |
torch._inductor.exc.LoweringException: RuntimeError: Overloaded torch operator invoked from Python failed to many any schema: | |
aten::arange() expected at most 5 argument(s) but received 7 argument(s). Declaration: aten::arange(Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() expected at most 6 argument(s) but received 7 argument(s). Declaration: aten::arange.start(Scalar start, Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() Expected a value of type 'number' for argument 'end' but instead found type 'Symbol'. | |
Position: 1 | |
Value: s1 | |
Declaration: aten::arange.start_step(Scalar start, Scalar end, Scalar step=1, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
Cast error details: Cannot cast s1 to number | |
aten::arange() expected at most 4 argument(s) but received 7 argument(s). Declaration: aten::arange.start_out(Scalar start, Scalar end, Scalar step=1, *, Tensor(a!) out) -> Tensor(a!) | |
aten::arange() expected at most 2 argument(s) but received 7 argument(s). Declaration: aten::arange.out(Scalar end, *, Tensor(a!) out) -> Tensor(a!) | |
target: aten.arange.default | |
args[0]: s1 | |
kwargs: {'dtype': torch.int64, 'device': device(type='cuda', index=0), 'pin_memory': False} | |
While executing %arange : [#users=1] = call_function[target=torch.ops.aten.arange.default](args = (%sym_size,), kwargs = {dtype: torch.int64, device: cuda:0, pin_memory: False}) | |
Original traceback: | |
Module stack: {'self_embeddings': <class 'fastNLP.modules.encoder.bert.BertEmbeddings'>} | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/fastNLP/modules/encoder/bert.py", line 230, in forward | |
position_ids = torch.arange(seq_length, dtype=torch.long, device=input_ids.device) | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/fastNLP/modules/encoder/bert.py", line 512, in forward | |
embedding_output = self.embeddings(input_ids, token_type_ids) | |
TorchDynamo optimized model failed to run because of following error | |
cuda train fastNLP_Bert FAIL | |
Running torchbench.py functorch_dp_cifar10... | |
cuda train functorch_dp_cifar10 FAIL (TIMEOUT) | |
Running torchbench.py functorch_maml_omniglot... | |
ERROR:common:'int' object has no attribute 'size' | |
While executing %sym_size : [#users=2] = placeholder[target=sym_size] | |
Original traceback: | |
Module stack: {'self_12': <class 'torch.nn.modules.flatten.Flatten'>} | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/container.py", line 204, in forward | |
input = module(input) | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1041, in check_accuracy | |
new_result = optimized_model_iter_fn(model, example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 945, in run_n_iterations | |
self.model_iter_fn(mod, inputs, collect_outputs=False) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 332, in forward_and_backward_pass | |
cloned_inputs = clone_inputs(inputs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 337, in <graph break in forward_and_backward_pass> | |
self.grad_scaler.scale(loss).backward() | |
File "/data/users/ezyang/pytorch-tmp/torch/_tensor.py", line 488, in backward | |
torch.autograd.backward( | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/__init__.py", line 197, in backward | |
Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/function.py", line 270, in apply | |
return user_fn(self, *args) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 499, in backward | |
CompiledFunction.compiled_bw = aot_config.bw_compiler( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/optimizations/backends.py", line 555, in _wrapped_bw_compiler | |
return disable(bw_compiler(*args, **kwargs)) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 362, in bw_compiler | |
return compile_fx_inner( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/debug_utils.py", line 466, in debug_wrapper | |
compiled_fn = compiler_fn(gm, example_inputs, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/debug.py", line 177, in inner | |
return fn(*args, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/contextlib.py", line 79, in inner | |
return func(*args, **kwds) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 122, in compile_fx_inner | |
graph.run(*example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 129, in run | |
return super().run(*args) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 130, in run | |
self.env[node] = self.run_node(node) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 299, in run_node | |
result = super().run_node(n) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 171, in run_node | |
return getattr(self, n.op)(n.target, args, kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 199, in placeholder | |
sizes, strides = self.static_sizes_strides(example) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 53, in static_sizes_strides | |
size = [sympy.Integer(i) for i in ex.size()] | |
AttributeError: 'int' object has no attribute 'size' | |
While executing %sym_size : [#users=2] = placeholder[target=sym_size] | |
Original traceback: | |
Module stack: {'self_12': <class 'torch.nn.modules.flatten.Flatten'>} | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/container.py", line 204, in forward | |
input = module(input) | |
TorchDynamo optimized model failed to run because of following error | |
cuda train functorch_maml_omniglot FAIL | |
Running torchbench.py hf_Albert... | |
ERROR:common:'int' object has no attribute 'size' | |
While executing %sym_size : [#users=2] = placeholder[target=sym_size] | |
Original traceback: | |
Module stack: {'self_predictions': <class 'transformers.models.albert.modeling_albert.AlbertMLMHead'>, 'self_predictions_dense': <class 'torch.nn.modules.linear.Linear'>} | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/albert/modeling_albert.py", line 883, in forward | |
hidden_states = self.dense(hidden_states) | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/albert/modeling_albert.py", line 1003, in <graph break in forward> | |
prediction_scores = self.predictions(sequence_outputs) | |
Gradient addition node due to multiple use of tensor around: | |
Module stack: {'self_embedding_hidden_mapping_in': <class 'torch.nn.modules.linear.Linear'>} | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/albert/modeling_albert.py", line 470, in forward | |
hidden_states = self.embedding_hidden_mapping_in(hidden_states) | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1041, in check_accuracy | |
new_result = optimized_model_iter_fn(model, example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 945, in run_n_iterations | |
self.model_iter_fn(mod, inputs, collect_outputs=False) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 332, in forward_and_backward_pass | |
cloned_inputs = clone_inputs(inputs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 337, in <graph break in forward_and_backward_pass> | |
self.grad_scaler.scale(loss).backward() | |
File "/data/users/ezyang/pytorch-tmp/torch/_tensor.py", line 488, in backward | |
torch.autograd.backward( | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/__init__.py", line 197, in backward | |
Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/function.py", line 270, in apply | |
return user_fn(self, *args) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 499, in backward | |
CompiledFunction.compiled_bw = aot_config.bw_compiler( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/optimizations/backends.py", line 555, in _wrapped_bw_compiler | |
return disable(bw_compiler(*args, **kwargs)) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 362, in bw_compiler | |
return compile_fx_inner( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/debug_utils.py", line 466, in debug_wrapper | |
compiled_fn = compiler_fn(gm, example_inputs, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/debug.py", line 177, in inner | |
return fn(*args, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/contextlib.py", line 79, in inner | |
return func(*args, **kwds) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 122, in compile_fx_inner | |
graph.run(*example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 129, in run | |
return super().run(*args) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 130, in run | |
self.env[node] = self.run_node(node) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 299, in run_node | |
result = super().run_node(n) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 171, in run_node | |
return getattr(self, n.op)(n.target, args, kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 199, in placeholder | |
sizes, strides = self.static_sizes_strides(example) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 53, in static_sizes_strides | |
size = [sympy.Integer(i) for i in ex.size()] | |
AttributeError: 'int' object has no attribute 'size' | |
While executing %sym_size : [#users=2] = placeholder[target=sym_size] | |
Original traceback: | |
Module stack: {'self_predictions': <class 'transformers.models.albert.modeling_albert.AlbertMLMHead'>, 'self_predictions_dense': <class 'torch.nn.modules.linear.Linear'>} | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/albert/modeling_albert.py", line 883, in forward | |
hidden_states = self.dense(hidden_states) | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/albert/modeling_albert.py", line 1003, in <graph break in forward> | |
prediction_scores = self.predictions(sequence_outputs) | |
Gradient addition node due to multiple use of tensor around: | |
Module stack: {'self_embedding_hidden_mapping_in': <class 'torch.nn.modules.linear.Linear'>} | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/albert/modeling_albert.py", line 470, in forward | |
hidden_states = self.embedding_hidden_mapping_in(hidden_states) | |
TorchDynamo optimized model failed to run because of following error | |
cuda train hf_Albert FAIL | |
Running torchbench.py hf_Bart... | |
ERROR:common:RuntimeError: Overloaded torch operator invoked from Python failed to many any schema: | |
aten::arange() expected at most 5 argument(s) but received 7 argument(s). Declaration: aten::arange(Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() expected at most 6 argument(s) but received 7 argument(s). Declaration: aten::arange.start(Scalar start, Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() Expected a value of type 'number' for argument 'end' but instead found type 'Symbol'. | |
Position: 1 | |
Value: s1 | |
Declaration: aten::arange.start_step(Scalar start, Scalar end, Scalar step=1, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
Cast error details: Cannot cast s1 to number | |
aten::arange() expected at most 4 argument(s) but received 7 argument(s). Declaration: aten::arange.start_out(Scalar start, Scalar end, Scalar step=1, *, Tensor(a!) out) -> Tensor(a!) | |
aten::arange() expected at most 2 argument(s) but received 7 argument(s). Declaration: aten::arange.out(Scalar end, *, Tensor(a!) out) -> Tensor(a!) | |
target: aten.arange.start | |
args[0]: 0 | |
args[1]: s1 | |
kwargs: {'dtype': torch.int64, 'device': device(type='cuda', index=0), 'pin_memory': False} | |
While executing %arange : [#users=1] = call_function[target=torch.ops.aten.arange.start](args = (0, %add), kwargs = {dtype: torch.int64, device: cuda:0, pin_memory: False}) | |
Original traceback: | |
Module stack: {'self_embed_positions': <class 'transformers.models.bart.modeling_bart.BartLearnedPositionalEmbedding'>} | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/bart/modeling_bart.py", line 134, in forward | |
positions = torch.arange( | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/bart/modeling_bart.py", line 801, in forward | |
embed_pos = self.embed_positions(input_shape) | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 239, in call_function | |
out = lowerings[target](*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/lowering.py", line 204, in wrapped | |
return decomp_fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/lowering.py", line 1194, in arange | |
return fallback_arange( | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/lowering.py", line 893, in handler | |
result = ir.FallbackKernel.create(kernel, *args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/ir.py", line 2890, in create | |
example_output = kernel( | |
File "/data/users/ezyang/pytorch-tmp/torch/_ops.py", line 446, in __call__ | |
return self._op(*args, **kwargs or {}) | |
RuntimeError: Overloaded torch operator invoked from Python failed to many any schema: | |
aten::arange() expected at most 5 argument(s) but received 7 argument(s). Declaration: aten::arange(Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() expected at most 6 argument(s) but received 7 argument(s). Declaration: aten::arange.start(Scalar start, Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() Expected a value of type 'number' for argument 'end' but instead found type 'Symbol'. | |
Position: 1 | |
Value: s1 | |
Declaration: aten::arange.start_step(Scalar start, Scalar end, Scalar step=1, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
Cast error details: Cannot cast s1 to number | |
aten::arange() expected at most 4 argument(s) but received 7 argument(s). Declaration: aten::arange.start_out(Scalar start, Scalar end, Scalar step=1, *, Tensor(a!) out) -> Tensor(a!) | |
aten::arange() expected at most 2 argument(s) but received 7 argument(s). Declaration: aten::arange.out(Scalar end, *, Tensor(a!) out) -> Tensor(a!) | |
The above exception was the direct cause of the following exception: | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1041, in check_accuracy | |
new_result = optimized_model_iter_fn(model, example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 945, in run_n_iterations | |
self.model_iter_fn(mod, inputs, collect_outputs=False) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 332, in forward_and_backward_pass | |
cloned_inputs = clone_inputs(inputs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 335, in <graph break in forward_and_backward_pass> | |
pred = mod(*cloned_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "/data/users/ezyang/benchmark/torchbenchmark/util/framework/huggingface/model_factory.py", line 41, in forward | |
return self.model(input_ids=input_ids, decoder_input_ids=decoder_input_ids) | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/bart/modeling_bart.py", line 1353, in forward | |
outputs = self.model( | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/bart/modeling_bart.py", line 1222, in forward | |
encoder_outputs = self.encoder( | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/bart/modeling_bart.py", line 735, in forward | |
def forward( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 893, in forward | |
return compiled_f( | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 880, in new_func | |
compiled_fn = create_aot_dispatcher_function( | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 600, in create_aot_dispatcher_function | |
aot_dispatch_autograd(flat_fn, fake_flat_tensor_args, aot_config) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 459, in aot_dispatch_autograd | |
compiled_fw_func = aot_config.fw_compiler(fw_module, deduped_flat_args) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 351, in fw_compiler | |
return compile_fx_inner( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/debug_utils.py", line 466, in debug_wrapper | |
compiled_fn = compiler_fn(gm, example_inputs, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/debug.py", line 177, in inner | |
return fn(*args, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/contextlib.py", line 79, in inner | |
return func(*args, **kwds) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 122, in compile_fx_inner | |
graph.run(*example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 129, in run | |
return super().run(*args) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 130, in run | |
self.env[node] = self.run_node(node) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 299, in run_node | |
result = super().run_node(n) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 171, in run_node | |
return getattr(self, n.op)(n.target, args, kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 242, in call_function | |
raise LoweringException(e, target, args, kwargs) from e | |
torch._inductor.exc.LoweringException: RuntimeError: Overloaded torch operator invoked from Python failed to many any schema: | |
aten::arange() expected at most 5 argument(s) but received 7 argument(s). Declaration: aten::arange(Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() expected at most 6 argument(s) but received 7 argument(s). Declaration: aten::arange.start(Scalar start, Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() Expected a value of type 'number' for argument 'end' but instead found type 'Symbol'. | |
Position: 1 | |
Value: s1 | |
Declaration: aten::arange.start_step(Scalar start, Scalar end, Scalar step=1, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
Cast error details: Cannot cast s1 to number | |
aten::arange() expected at most 4 argument(s) but received 7 argument(s). Declaration: aten::arange.start_out(Scalar start, Scalar end, Scalar step=1, *, Tensor(a!) out) -> Tensor(a!) | |
aten::arange() expected at most 2 argument(s) but received 7 argument(s). Declaration: aten::arange.out(Scalar end, *, Tensor(a!) out) -> Tensor(a!) | |
target: aten.arange.start | |
args[0]: 0 | |
args[1]: s1 | |
kwargs: {'dtype': torch.int64, 'device': device(type='cuda', index=0), 'pin_memory': False} | |
While executing %arange : [#users=1] = call_function[target=torch.ops.aten.arange.start](args = (0, %add), kwargs = {dtype: torch.int64, device: cuda:0, pin_memory: False}) | |
Original traceback: | |
Module stack: {'self_embed_positions': <class 'transformers.models.bart.modeling_bart.BartLearnedPositionalEmbedding'>} | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/bart/modeling_bart.py", line 134, in forward | |
positions = torch.arange( | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/bart/modeling_bart.py", line 801, in forward | |
embed_pos = self.embed_positions(input_shape) | |
TorchDynamo optimized model failed to run because of following error | |
cuda train hf_Bart FAIL | |
Running torchbench.py hf_Bert... | |
ERROR:common:'int' object has no attribute 'size' | |
While executing %sym_size : [#users=2] = placeholder[target=sym_size] | |
Original traceback: | |
Module stack: {'self_cls': <class 'transformers.models.bert.modeling_bert.BertOnlyMLMHead'>, 'self_cls_predictions': <class 'transformers.models.bert.modeling_bert.BertLMPredictionHead'>, 'self_cls_predictions_transform': <class 'transformers.models.bert.modeling_bert.BertPredictionHeadTransform'>, 'self_cls_predictions_transform_dense': <class 'torch.nn.modules.linear.Linear'>} | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/bert/modeling_bert.py", line 675, in forward | |
hidden_states = self.dense(hidden_states) | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/bert/modeling_bert.py", line 696, in forward | |
hidden_states = self.transform(hidden_states) | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/bert/modeling_bert.py", line 707, in forward | |
prediction_scores = self.predictions(sequence_output) | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/bert/modeling_bert.py", line 1366, in <graph break in forward> | |
prediction_scores = self.cls(sequence_output) | |
Gradient addition node due to multiple use of tensor around: | |
Module stack: {'self_layer_0': <class 'transformers.models.bert.modeling_bert.BertLayer'>, 'self_layer_0_attention': <class 'transformers.models.bert.modeling_bert.BertAttention'>, 'self_layer_0_attention_self': <class 'transformers.models.bert.modeling_bert.BertSelfAttention'>, 'self_layer_0_attention_self_query': <class 'torch.nn.modules.linear.Linear'>} | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/bert/modeling_bert.py", line 289, in forward | |
mixed_query_layer = self.query(hidden_states) | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/bert/modeling_bert.py", line 423, in forward | |
self_outputs = self.self( | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/bert/modeling_bert.py", line 493, in forward | |
self_attention_outputs = self.attention( | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/bert/modeling_bert.py", line 607, in forward | |
layer_outputs = layer_module( | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1041, in check_accuracy | |
new_result = optimized_model_iter_fn(model, example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 945, in run_n_iterations | |
self.model_iter_fn(mod, inputs, collect_outputs=False) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 332, in forward_and_backward_pass | |
cloned_inputs = clone_inputs(inputs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 337, in <graph break in forward_and_backward_pass> | |
self.grad_scaler.scale(loss).backward() | |
File "/data/users/ezyang/pytorch-tmp/torch/_tensor.py", line 488, in backward | |
torch.autograd.backward( | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/__init__.py", line 197, in backward | |
Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/function.py", line 270, in apply | |
return user_fn(self, *args) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 499, in backward | |
CompiledFunction.compiled_bw = aot_config.bw_compiler( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/optimizations/backends.py", line 555, in _wrapped_bw_compiler | |
return disable(bw_compiler(*args, **kwargs)) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 362, in bw_compiler | |
return compile_fx_inner( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/debug_utils.py", line 466, in debug_wrapper | |
compiled_fn = compiler_fn(gm, example_inputs, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/debug.py", line 177, in inner | |
return fn(*args, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/contextlib.py", line 79, in inner | |
return func(*args, **kwds) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 122, in compile_fx_inner | |
graph.run(*example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 129, in run | |
return super().run(*args) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 130, in run | |
self.env[node] = self.run_node(node) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 299, in run_node | |
result = super().run_node(n) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 171, in run_node | |
return getattr(self, n.op)(n.target, args, kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 199, in placeholder | |
sizes, strides = self.static_sizes_strides(example) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 53, in static_sizes_strides | |
size = [sympy.Integer(i) for i in ex.size()] | |
AttributeError: 'int' object has no attribute 'size' | |
While executing %sym_size : [#users=2] = placeholder[target=sym_size] | |
Original traceback: | |
Module stack: {'self_cls': <class 'transformers.models.bert.modeling_bert.BertOnlyMLMHead'>, 'self_cls_predictions': <class 'transformers.models.bert.modeling_bert.BertLMPredictionHead'>, 'self_cls_predictions_transform': <class 'transformers.models.bert.modeling_bert.BertPredictionHeadTransform'>, 'self_cls_predictions_transform_dense': <class 'torch.nn.modules.linear.Linear'>} | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/bert/modeling_bert.py", line 675, in forward | |
hidden_states = self.dense(hidden_states) | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/bert/modeling_bert.py", line 696, in forward | |
hidden_states = self.transform(hidden_states) | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/bert/modeling_bert.py", line 707, in forward | |
prediction_scores = self.predictions(sequence_output) | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/bert/modeling_bert.py", line 1366, in <graph break in forward> | |
prediction_scores = self.cls(sequence_output) | |
Gradient addition node due to multiple use of tensor around: | |
Module stack: {'self_layer_0': <class 'transformers.models.bert.modeling_bert.BertLayer'>, 'self_layer_0_attention': <class 'transformers.models.bert.modeling_bert.BertAttention'>, 'self_layer_0_attention_self': <class 'transformers.models.bert.modeling_bert.BertSelfAttention'>, 'self_layer_0_attention_self_query': <class 'torch.nn.modules.linear.Linear'>} | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/bert/modeling_bert.py", line 289, in forward | |
mixed_query_layer = self.query(hidden_states) | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/bert/modeling_bert.py", line 423, in forward | |
self_outputs = self.self( | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/bert/modeling_bert.py", line 493, in forward | |
self_attention_outputs = self.attention( | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/bert/modeling_bert.py", line 607, in forward | |
layer_outputs = layer_module( | |
TorchDynamo optimized model failed to run because of following error | |
cuda train hf_Bert FAIL | |
Running torchbench.py hf_BigBird... | |
ERROR:common:(False, s0, s1) | |
While executing return (False, sym_size, sym_size_1) | |
Original traceback: | |
None | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1041, in check_accuracy | |
new_result = optimized_model_iter_fn(model, example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 945, in run_n_iterations | |
self.model_iter_fn(mod, inputs, collect_outputs=False) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 332, in forward_and_backward_pass | |
cloned_inputs = clone_inputs(inputs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 335, in <graph break in forward_and_backward_pass> | |
pred = mod(*cloned_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/big_bird/modeling_big_bird.py", line 2462, in forward | |
outputs = self.bert( | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/big_bird/modeling_big_bird.py", line 2104, in forward | |
blocked_encoder_mask, band_mask, from_mask, to_mask = self.create_masks_for_block_sparse_attn( | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/big_bird/modeling_big_bird.py", line 2185, in create_masks_for_block_sparse_attn | |
@staticmethod | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 893, in forward | |
return compiled_f( | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 880, in new_func | |
compiled_fn = create_aot_dispatcher_function( | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 603, in create_aot_dispatcher_function | |
return aot_dispatch_base(flat_fn, fake_flat_tensor_args, aot_config) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 336, in aot_dispatch_base | |
compiled_fw = aot_config.fw_compiler(fw_module, flat_args) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 351, in fw_compiler | |
return compile_fx_inner( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/debug_utils.py", line 466, in debug_wrapper | |
compiled_fn = compiler_fn(gm, example_inputs, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/debug.py", line 177, in inner | |
return fn(*args, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/contextlib.py", line 79, in inner | |
return func(*args, **kwds) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 122, in compile_fx_inner | |
graph.run(*example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 129, in run | |
return super().run(*args) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 130, in run | |
self.env[node] = self.run_node(node) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 299, in run_node | |
result = super().run_node(n) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 171, in run_node | |
return getattr(self, n.op)(n.target, args, kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 267, in output | |
assert all( | |
AssertionError: (False, s0, s1) | |
While executing return (False, sym_size, sym_size_1) | |
Original traceback: | |
None | |
TorchDynamo optimized model failed to run because of following error | |
cuda train hf_BigBird FAIL | |
Running torchbench.py hf_DistilBert... | |
ERROR:common:'int' object has no attribute 'size' | |
While executing %sym_size : [#users=2] = placeholder[target=sym_size] | |
Original traceback: | |
Module stack: {'self_vocab_transform': <class 'torch.nn.modules.linear.Linear'>} | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/distilbert/modeling_distilbert.py", line 659, in <graph break in forward> | |
prediction_logits = self.vocab_transform(hidden_states) # (bs, seq_length, dim) | |
Gradient addition node due to multiple use of tensor around: | |
Module stack: {'self_transformer': <class 'transformers.models.distilbert.modeling_distilbert.Transformer'>, 'self_transformer_layer_0': <class 'transformers.models.distilbert.modeling_distilbert.TransformerBlock'>, 'self_transformer_layer_0_attention': <class 'transformers.models.distilbert.modeling_distilbert.MultiHeadSelfAttention'>, 'self_transformer_layer_0_attention_q_lin': <class 'torch.nn.modules.linear.Linear'>} | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/distilbert/modeling_distilbert.py", line 207, in forward | |
q = shape(self.q_lin(query)) # (bs, n_heads, q_length, dim_per_head) | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/distilbert/modeling_distilbert.py", line 283, in forward | |
sa_output = self.attention( | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/distilbert/modeling_distilbert.py", line 345, in forward | |
layer_outputs = layer_module( | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/distilbert/modeling_distilbert.py", line 567, in <graph break in forward> | |
return self.transformer( | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1041, in check_accuracy | |
new_result = optimized_model_iter_fn(model, example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 945, in run_n_iterations | |
self.model_iter_fn(mod, inputs, collect_outputs=False) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 332, in forward_and_backward_pass | |
cloned_inputs = clone_inputs(inputs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 337, in <graph break in forward_and_backward_pass> | |
self.grad_scaler.scale(loss).backward() | |
File "/data/users/ezyang/pytorch-tmp/torch/_tensor.py", line 488, in backward | |
torch.autograd.backward( | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/__init__.py", line 197, in backward | |
Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/function.py", line 270, in apply | |
return user_fn(self, *args) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 499, in backward | |
CompiledFunction.compiled_bw = aot_config.bw_compiler( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/optimizations/backends.py", line 555, in _wrapped_bw_compiler | |
return disable(bw_compiler(*args, **kwargs)) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 362, in bw_compiler | |
return compile_fx_inner( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/debug_utils.py", line 466, in debug_wrapper | |
compiled_fn = compiler_fn(gm, example_inputs, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/debug.py", line 177, in inner | |
return fn(*args, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/contextlib.py", line 79, in inner | |
return func(*args, **kwds) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 122, in compile_fx_inner | |
graph.run(*example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 129, in run | |
return super().run(*args) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 130, in run | |
self.env[node] = self.run_node(node) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 299, in run_node | |
result = super().run_node(n) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 171, in run_node | |
return getattr(self, n.op)(n.target, args, kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 199, in placeholder | |
sizes, strides = self.static_sizes_strides(example) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 53, in static_sizes_strides | |
size = [sympy.Integer(i) for i in ex.size()] | |
AttributeError: 'int' object has no attribute 'size' | |
While executing %sym_size : [#users=2] = placeholder[target=sym_size] | |
Original traceback: | |
Module stack: {'self_vocab_transform': <class 'torch.nn.modules.linear.Linear'>} | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/distilbert/modeling_distilbert.py", line 659, in <graph break in forward> | |
prediction_logits = self.vocab_transform(hidden_states) # (bs, seq_length, dim) | |
Gradient addition node due to multiple use of tensor around: | |
Module stack: {'self_transformer': <class 'transformers.models.distilbert.modeling_distilbert.Transformer'>, 'self_transformer_layer_0': <class 'transformers.models.distilbert.modeling_distilbert.TransformerBlock'>, 'self_transformer_layer_0_attention': <class 'transformers.models.distilbert.modeling_distilbert.MultiHeadSelfAttention'>, 'self_transformer_layer_0_attention_q_lin': <class 'torch.nn.modules.linear.Linear'>} | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/distilbert/modeling_distilbert.py", line 207, in forward | |
q = shape(self.q_lin(query)) # (bs, n_heads, q_length, dim_per_head) | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/distilbert/modeling_distilbert.py", line 283, in forward | |
sa_output = self.attention( | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/distilbert/modeling_distilbert.py", line 345, in forward | |
layer_outputs = layer_module( | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/distilbert/modeling_distilbert.py", line 567, in <graph break in forward> | |
return self.transformer( | |
TorchDynamo optimized model failed to run because of following error | |
cuda train hf_DistilBert FAIL | |
Running torchbench.py hf_GPT2... | |
ERROR:common:RuntimeError: Overloaded torch operator invoked from Python failed to many any schema: | |
aten::arange() expected at most 5 argument(s) but received 7 argument(s). Declaration: aten::arange(Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() expected at most 6 argument(s) but received 7 argument(s). Declaration: aten::arange.start(Scalar start, Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() Expected a value of type 'number' for argument 'end' but instead found type 'Symbol'. | |
Position: 1 | |
Value: s1 | |
Declaration: aten::arange.start_step(Scalar start, Scalar end, Scalar step=1, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
Cast error details: Cannot cast s1 to number | |
aten::arange() expected at most 4 argument(s) but received 7 argument(s). Declaration: aten::arange.start_out(Scalar start, Scalar end, Scalar step=1, *, Tensor(a!) out) -> Tensor(a!) | |
aten::arange() expected at most 2 argument(s) but received 7 argument(s). Declaration: aten::arange.out(Scalar end, *, Tensor(a!) out) -> Tensor(a!) | |
target: aten.arange.start | |
args[0]: 0 | |
args[1]: s1 | |
kwargs: {'dtype': torch.int64, 'device': device(type='cuda', index=0), 'pin_memory': False} | |
While executing %arange : [#users=1] = call_function[target=torch.ops.aten.arange.start](args = (0, %add), kwargs = {dtype: torch.int64, device: cuda:0, pin_memory: False}) | |
Original traceback: | |
Module stack: {} | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/gpt2/modeling_gpt2.py", line 793, in forward | |
position_ids = torch.arange(past_length, input_shape[-1] + past_length, dtype=torch.long, device=device) | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 239, in call_function | |
out = lowerings[target](*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/lowering.py", line 204, in wrapped | |
return decomp_fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/lowering.py", line 1194, in arange | |
return fallback_arange( | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/lowering.py", line 893, in handler | |
result = ir.FallbackKernel.create(kernel, *args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/ir.py", line 2890, in create | |
example_output = kernel( | |
File "/data/users/ezyang/pytorch-tmp/torch/_ops.py", line 446, in __call__ | |
return self._op(*args, **kwargs or {}) | |
RuntimeError: Overloaded torch operator invoked from Python failed to many any schema: | |
aten::arange() expected at most 5 argument(s) but received 7 argument(s). Declaration: aten::arange(Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() expected at most 6 argument(s) but received 7 argument(s). Declaration: aten::arange.start(Scalar start, Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() Expected a value of type 'number' for argument 'end' but instead found type 'Symbol'. | |
Position: 1 | |
Value: s1 | |
Declaration: aten::arange.start_step(Scalar start, Scalar end, Scalar step=1, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
Cast error details: Cannot cast s1 to number | |
aten::arange() expected at most 4 argument(s) but received 7 argument(s). Declaration: aten::arange.start_out(Scalar start, Scalar end, Scalar step=1, *, Tensor(a!) out) -> Tensor(a!) | |
aten::arange() expected at most 2 argument(s) but received 7 argument(s). Declaration: aten::arange.out(Scalar end, *, Tensor(a!) out) -> Tensor(a!) | |
The above exception was the direct cause of the following exception: | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1041, in check_accuracy | |
new_result = optimized_model_iter_fn(model, example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 945, in run_n_iterations | |
self.model_iter_fn(mod, inputs, collect_outputs=False) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 332, in forward_and_backward_pass | |
cloned_inputs = clone_inputs(inputs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 335, in <graph break in forward_and_backward_pass> | |
pred = mod(*cloned_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/gpt2/modeling_gpt2.py", line 1048, in forward | |
transformer_outputs = self.transformer( | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/gpt2/modeling_gpt2.py", line 738, in forward | |
@add_start_docstrings_to_model_forward(GPT2_INPUTS_DOCSTRING) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 893, in forward | |
return compiled_f( | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 880, in new_func | |
compiled_fn = create_aot_dispatcher_function( | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 600, in create_aot_dispatcher_function | |
aot_dispatch_autograd(flat_fn, fake_flat_tensor_args, aot_config) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 459, in aot_dispatch_autograd | |
compiled_fw_func = aot_config.fw_compiler(fw_module, deduped_flat_args) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 351, in fw_compiler | |
return compile_fx_inner( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/debug_utils.py", line 466, in debug_wrapper | |
compiled_fn = compiler_fn(gm, example_inputs, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/debug.py", line 177, in inner | |
return fn(*args, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/contextlib.py", line 79, in inner | |
return func(*args, **kwds) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 122, in compile_fx_inner | |
graph.run(*example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 129, in run | |
return super().run(*args) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 130, in run | |
self.env[node] = self.run_node(node) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 299, in run_node | |
result = super().run_node(n) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 171, in run_node | |
return getattr(self, n.op)(n.target, args, kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 242, in call_function | |
raise LoweringException(e, target, args, kwargs) from e | |
torch._inductor.exc.LoweringException: RuntimeError: Overloaded torch operator invoked from Python failed to many any schema: | |
aten::arange() expected at most 5 argument(s) but received 7 argument(s). Declaration: aten::arange(Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() expected at most 6 argument(s) but received 7 argument(s). Declaration: aten::arange.start(Scalar start, Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() Expected a value of type 'number' for argument 'end' but instead found type 'Symbol'. | |
Position: 1 | |
Value: s1 | |
Declaration: aten::arange.start_step(Scalar start, Scalar end, Scalar step=1, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
Cast error details: Cannot cast s1 to number | |
aten::arange() expected at most 4 argument(s) but received 7 argument(s). Declaration: aten::arange.start_out(Scalar start, Scalar end, Scalar step=1, *, Tensor(a!) out) -> Tensor(a!) | |
aten::arange() expected at most 2 argument(s) but received 7 argument(s). Declaration: aten::arange.out(Scalar end, *, Tensor(a!) out) -> Tensor(a!) | |
target: aten.arange.start | |
args[0]: 0 | |
args[1]: s1 | |
kwargs: {'dtype': torch.int64, 'device': device(type='cuda', index=0), 'pin_memory': False} | |
While executing %arange : [#users=1] = call_function[target=torch.ops.aten.arange.start](args = (0, %add), kwargs = {dtype: torch.int64, device: cuda:0, pin_memory: False}) | |
Original traceback: | |
Module stack: {} | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/gpt2/modeling_gpt2.py", line 793, in forward | |
position_ids = torch.arange(past_length, input_shape[-1] + past_length, dtype=torch.long, device=device) | |
TorchDynamo optimized model failed to run because of following error | |
cuda train hf_GPT2 FAIL | |
Running torchbench.py hf_GPT2_large... | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 349, in <module> | |
main(TorchBenchmarkRunner(), original_dir) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1775, in main | |
runner.run_one_model( | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 768, in inner | |
return fn(self, *args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1213, in run_one_model | |
status = self.check_accuracy( | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1023, in check_accuracy | |
correct_rerun_result = self.run_n_iterations( | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 946, in run_n_iterations | |
return self.model_iter_fn(mod, inputs, collect_outputs=True) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 335, in forward_and_backward_pass | |
pred = mod(*cloned_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/gpt2/modeling_gpt2.py", line 1048, in forward | |
transformer_outputs = self.transformer( | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/gpt2/modeling_gpt2.py", line 891, in forward | |
outputs = block( | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/gpt2/modeling_gpt2.py", line 428, in forward | |
feed_forward_hidden_states = self.mlp(hidden_states) | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/gpt2/modeling_gpt2.py", line 356, in forward | |
hidden_states = self.act(hidden_states) | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/activations.py", line 34, in forward | |
return 0.5 * input * (1.0 + torch.tanh(math.sqrt(2.0 / math.pi) * (input + 0.044715 * torch.pow(input, 3.0)))) | |
torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 20.00 MiB (GPU 0; 39.59 GiB total capacity; 37.54 GiB already allocated; 1.44 MiB free; 38.41 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF | |
cuda train hf_GPT2_large FAIL | |
Running torchbench.py hf_Longformer... | |
[2022-10-30 21:21:38,092] torch._dynamo.variables.builtin: [WARNING] incorrect arg count <bound method BuiltinVariable._call_min_max of BuiltinVariable(max)> missing a required argument: 'b' | |
[2022-10-30 21:21:38,122] torch._dynamo.variables.builtin: [WARNING] incorrect arg count <bound method BuiltinVariable._call_min_max of BuiltinVariable(max)> missing a required argument: 'b' | |
[2022-10-30 21:21:39,013] torch._dynamo.variables.builtin: [WARNING] incorrect arg count <bound method BuiltinVariable._call_min_max of BuiltinVariable(max)> missing a required argument: 'b' | |
[2022-10-30 21:21:42,296] torch._inductor.ir: [WARNING] Using FallbackKernel: aten.cumsum | |
[2022-10-30 21:21:44,860] torch._inductor.graph: [WARNING] Creating implicit fallback for: | |
target: <built-in function eq> | |
args[0]: 768 | |
args[1]: 768 | |
ERROR:common:AttributeError: 'bool' object has no attribute 'device' | |
target: <built-in function eq> | |
args[0]: 768 | |
args[1]: 768 | |
While executing %eq : [#users=1] = call_function[target=operator.eq](args = (%sym_size_2, 768), kwargs = {}) | |
Original traceback: | |
None | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 239, in call_function | |
out = lowerings[target](*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/lowering.py", line 204, in wrapped | |
return decomp_fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/lowering.py", line 893, in handler | |
result = ir.FallbackKernel.create(kernel, *args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/ir.py", line 2922, in create | |
example_output.device, | |
AttributeError: 'bool' object has no attribute 'device' | |
The above exception was the direct cause of the following exception: | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1041, in check_accuracy | |
new_result = optimized_model_iter_fn(model, example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 945, in run_n_iterations | |
self.model_iter_fn(mod, inputs, collect_outputs=False) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 332, in forward_and_backward_pass | |
cloned_inputs = clone_inputs(inputs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 335, in <graph break in forward_and_backward_pass> | |
pred = mod(*cloned_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/longformer/modeling_longformer.py", line 1813, in forward | |
outputs = self.longformer( | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/longformer/modeling_longformer.py", line 1696, in forward | |
padding_len, input_ids, attention_mask, token_type_ids, position_ids, inputs_embeds = self._pad_to_window_size( | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/longformer/modeling_longformer.py", line 1715, in <graph break in forward> | |
encoder_outputs = self.encoder( | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/longformer/modeling_longformer.py", line 1265, in forward | |
is_global_attn = is_index_global_attn.flatten().any().item() | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/longformer/modeling_longformer.py", line 1297, in <graph break in forward> | |
layer_outputs = layer_module( | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/longformer/modeling_longformer.py", line 1221, in forward | |
self_attn_outputs = self.attention( | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/longformer/modeling_longformer.py", line 1157, in forward | |
self_outputs = self.self( | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/longformer/modeling_longformer.py", line 542, in forward | |
def forward( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 893, in forward | |
return compiled_f( | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 880, in new_func | |
compiled_fn = create_aot_dispatcher_function( | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 600, in create_aot_dispatcher_function | |
aot_dispatch_autograd(flat_fn, fake_flat_tensor_args, aot_config) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 459, in aot_dispatch_autograd | |
compiled_fw_func = aot_config.fw_compiler(fw_module, deduped_flat_args) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 351, in fw_compiler | |
return compile_fx_inner( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/debug_utils.py", line 466, in debug_wrapper | |
compiled_fn = compiler_fn(gm, example_inputs, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/debug.py", line 177, in inner | |
return fn(*args, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/contextlib.py", line 79, in inner | |
return func(*args, **kwds) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 122, in compile_fx_inner | |
graph.run(*example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 129, in run | |
return super().run(*args) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 130, in run | |
self.env[node] = self.run_node(node) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 299, in run_node | |
result = super().run_node(n) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 171, in run_node | |
return getattr(self, n.op)(n.target, args, kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 242, in call_function | |
raise LoweringException(e, target, args, kwargs) from e | |
torch._inductor.exc.LoweringException: AttributeError: 'bool' object has no attribute 'device' | |
target: <built-in function eq> | |
args[0]: 768 | |
args[1]: 768 | |
While executing %eq : [#users=1] = call_function[target=operator.eq](args = (%sym_size_2, 768), kwargs = {}) | |
Original traceback: | |
None | |
TorchDynamo optimized model failed to run because of following error | |
cuda train hf_Longformer FAIL | |
Running torchbench.py hf_Reformer... | |
ERROR:common:RuntimeError: Overloaded torch operator invoked from Python failed to many any schema: | |
aten::arange() expected at most 5 argument(s) but received 7 argument(s). Declaration: aten::arange(Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() expected at most 6 argument(s) but received 7 argument(s). Declaration: aten::arange.start(Scalar start, Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() Expected a value of type 'number' for argument 'end' but instead found type 'Symbol'. | |
Position: 1 | |
Value: s1 | |
Declaration: aten::arange.start_step(Scalar start, Scalar end, Scalar step=1, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
Cast error details: Cannot cast s1 to number | |
aten::arange() expected at most 4 argument(s) but received 7 argument(s). Declaration: aten::arange.start_out(Scalar start, Scalar end, Scalar step=1, *, Tensor(a!) out) -> Tensor(a!) | |
aten::arange() expected at most 2 argument(s) but received 7 argument(s). Declaration: aten::arange.out(Scalar end, *, Tensor(a!) out) -> Tensor(a!) | |
target: aten.arange.start | |
args[0]: 0 | |
args[1]: s1 | |
kwargs: {'dtype': torch.int64, 'device': device(type='cuda', index=0), 'pin_memory': False} | |
While executing %arange : [#users=1] = call_function[target=torch.ops.aten.arange.start](args = (0, %add), kwargs = {dtype: torch.int64, device: cuda:0, pin_memory: False}) | |
Original traceback: | |
Module stack: {} | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/reformer/modeling_reformer.py", line 249, in forward | |
position_ids = torch.arange( | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 239, in call_function | |
out = lowerings[target](*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/lowering.py", line 204, in wrapped | |
return decomp_fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/lowering.py", line 1194, in arange | |
return fallback_arange( | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/lowering.py", line 893, in handler | |
result = ir.FallbackKernel.create(kernel, *args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/ir.py", line 2890, in create | |
example_output = kernel( | |
File "/data/users/ezyang/pytorch-tmp/torch/_ops.py", line 446, in __call__ | |
return self._op(*args, **kwargs or {}) | |
RuntimeError: Overloaded torch operator invoked from Python failed to many any schema: | |
aten::arange() expected at most 5 argument(s) but received 7 argument(s). Declaration: aten::arange(Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() expected at most 6 argument(s) but received 7 argument(s). Declaration: aten::arange.start(Scalar start, Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() Expected a value of type 'number' for argument 'end' but instead found type 'Symbol'. | |
Position: 1 | |
Value: s1 | |
Declaration: aten::arange.start_step(Scalar start, Scalar end, Scalar step=1, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
Cast error details: Cannot cast s1 to number | |
aten::arange() expected at most 4 argument(s) but received 7 argument(s). Declaration: aten::arange.start_out(Scalar start, Scalar end, Scalar step=1, *, Tensor(a!) out) -> Tensor(a!) | |
aten::arange() expected at most 2 argument(s) but received 7 argument(s). Declaration: aten::arange.out(Scalar end, *, Tensor(a!) out) -> Tensor(a!) | |
The above exception was the direct cause of the following exception: | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1041, in check_accuracy | |
new_result = optimized_model_iter_fn(model, example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 945, in run_n_iterations | |
self.model_iter_fn(mod, inputs, collect_outputs=False) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 332, in forward_and_backward_pass | |
cloned_inputs = clone_inputs(inputs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 335, in <graph break in forward_and_backward_pass> | |
pred = mod(*cloned_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/reformer/modeling_reformer.py", line 2397, in forward | |
reformer_outputs = self.reformer( | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/reformer/modeling_reformer.py", line 2063, in forward | |
least_common_mult_chunk_length = _get_least_common_mult_chunk_len(self.config) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/reformer/modeling_reformer.py", line 2100, in <graph break in forward> | |
embedding_output = self.embeddings( | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/reformer/modeling_reformer.py", line 239, in forward | |
def forward(self, input_ids=None, position_ids=None, inputs_embeds=None, start_idx_pos_encodings=0): | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 893, in forward | |
return compiled_f( | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 880, in new_func | |
compiled_fn = create_aot_dispatcher_function( | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 600, in create_aot_dispatcher_function | |
aot_dispatch_autograd(flat_fn, fake_flat_tensor_args, aot_config) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 459, in aot_dispatch_autograd | |
compiled_fw_func = aot_config.fw_compiler(fw_module, deduped_flat_args) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 351, in fw_compiler | |
return compile_fx_inner( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/debug_utils.py", line 466, in debug_wrapper | |
compiled_fn = compiler_fn(gm, example_inputs, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/debug.py", line 177, in inner | |
return fn(*args, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/contextlib.py", line 79, in inner | |
return func(*args, **kwds) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 122, in compile_fx_inner | |
graph.run(*example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 129, in run | |
return super().run(*args) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 130, in run | |
self.env[node] = self.run_node(node) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 299, in run_node | |
result = super().run_node(n) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 171, in run_node | |
return getattr(self, n.op)(n.target, args, kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 242, in call_function | |
raise LoweringException(e, target, args, kwargs) from e | |
torch._inductor.exc.LoweringException: RuntimeError: Overloaded torch operator invoked from Python failed to many any schema: | |
aten::arange() expected at most 5 argument(s) but received 7 argument(s). Declaration: aten::arange(Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() expected at most 6 argument(s) but received 7 argument(s). Declaration: aten::arange.start(Scalar start, Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() Expected a value of type 'number' for argument 'end' but instead found type 'Symbol'. | |
Position: 1 | |
Value: s1 | |
Declaration: aten::arange.start_step(Scalar start, Scalar end, Scalar step=1, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
Cast error details: Cannot cast s1 to number | |
aten::arange() expected at most 4 argument(s) but received 7 argument(s). Declaration: aten::arange.start_out(Scalar start, Scalar end, Scalar step=1, *, Tensor(a!) out) -> Tensor(a!) | |
aten::arange() expected at most 2 argument(s) but received 7 argument(s). Declaration: aten::arange.out(Scalar end, *, Tensor(a!) out) -> Tensor(a!) | |
target: aten.arange.start | |
args[0]: 0 | |
args[1]: s1 | |
kwargs: {'dtype': torch.int64, 'device': device(type='cuda', index=0), 'pin_memory': False} | |
While executing %arange : [#users=1] = call_function[target=torch.ops.aten.arange.start](args = (0, %add), kwargs = {dtype: torch.int64, device: cuda:0, pin_memory: False}) | |
Original traceback: | |
Module stack: {} | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/reformer/modeling_reformer.py", line 249, in forward | |
position_ids = torch.arange( | |
TorchDynamo optimized model failed to run because of following error | |
cuda train hf_Reformer FAIL | |
Running torchbench.py hf_T5... | |
WARNING:common:fp64 golden ref were not generated for hf_T5 | |
ERROR:common:RuntimeError: Overloaded torch operator invoked from Python failed to many any schema: | |
aten::arange() expected at most 5 argument(s) but received 7 argument(s). Declaration: aten::arange(Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() expected at most 6 argument(s) but received 7 argument(s). Declaration: aten::arange.start(Scalar start, Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() Expected a value of type 'number' for argument 'end' but instead found type 'Symbol'. | |
Position: 1 | |
Value: s1 | |
Declaration: aten::arange.start_step(Scalar start, Scalar end, Scalar step=1, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
Cast error details: Cannot cast s1 to number | |
aten::arange() expected at most 4 argument(s) but received 7 argument(s). Declaration: aten::arange.start_out(Scalar start, Scalar end, Scalar step=1, *, Tensor(a!) out) -> Tensor(a!) | |
aten::arange() expected at most 2 argument(s) but received 7 argument(s). Declaration: aten::arange.out(Scalar end, *, Tensor(a!) out) -> Tensor(a!) | |
target: aten.arange.default | |
args[0]: s1 | |
kwargs: {'dtype': torch.int64, 'device': device(type='cuda', index=0), 'pin_memory': False} | |
While executing %arange : [#users=2] = call_function[target=torch.ops.aten.arange.default](args = (%sym_size_2,), kwargs = {dtype: torch.int64, device: cuda:0, pin_memory: False}) | |
Original traceback: | |
Module stack: {'self_encoder': <class 'transformers.models.t5.modeling_t5.T5Stack'>, 'self_encoder_block_0': <class 'transformers.models.t5.modeling_t5.T5Block'>, 'sub0_0': <class 'transformers.models.t5.modeling_t5.T5LayerSelfAttention'>, 'self_encoder_block_0_layer_0_SelfAttention': <class 'transformers.models.t5.modeling_t5.T5Attention'>} | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/t5/modeling_t5.py", line 423, in compute_bias | |
context_position = torch.arange(query_length, dtype=torch.long, device=device)[:, None] | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/t5/modeling_t5.py", line 519, in forward | |
position_bias = self.compute_bias(real_seq_length, key_length, device=scores.device) | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/t5/modeling_t5.py", line 570, in forward | |
attention_output = self.SelfAttention( | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/t5/modeling_t5.py", line 664, in forward | |
self_attention_outputs = self.layer[0]( | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/t5/modeling_t5.py", line 1033, in forward | |
layer_outputs = layer_module( | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/t5/modeling_t5.py", line 1601, in forward | |
encoder_outputs = self.encoder( | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 239, in call_function | |
out = lowerings[target](*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/lowering.py", line 204, in wrapped | |
return decomp_fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/lowering.py", line 1194, in arange | |
return fallback_arange( | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/lowering.py", line 893, in handler | |
result = ir.FallbackKernel.create(kernel, *args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/ir.py", line 2890, in create | |
example_output = kernel( | |
File "/data/users/ezyang/pytorch-tmp/torch/_ops.py", line 446, in __call__ | |
return self._op(*args, **kwargs or {}) | |
RuntimeError: Overloaded torch operator invoked from Python failed to many any schema: | |
aten::arange() expected at most 5 argument(s) but received 7 argument(s). Declaration: aten::arange(Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() expected at most 6 argument(s) but received 7 argument(s). Declaration: aten::arange.start(Scalar start, Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() Expected a value of type 'number' for argument 'end' but instead found type 'Symbol'. | |
Position: 1 | |
Value: s1 | |
Declaration: aten::arange.start_step(Scalar start, Scalar end, Scalar step=1, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
Cast error details: Cannot cast s1 to number | |
aten::arange() expected at most 4 argument(s) but received 7 argument(s). Declaration: aten::arange.start_out(Scalar start, Scalar end, Scalar step=1, *, Tensor(a!) out) -> Tensor(a!) | |
aten::arange() expected at most 2 argument(s) but received 7 argument(s). Declaration: aten::arange.out(Scalar end, *, Tensor(a!) out) -> Tensor(a!) | |
The above exception was the direct cause of the following exception: | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1041, in check_accuracy | |
new_result = optimized_model_iter_fn(model, example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 945, in run_n_iterations | |
self.model_iter_fn(mod, inputs, collect_outputs=False) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 332, in forward_and_backward_pass | |
cloned_inputs = clone_inputs(inputs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 335, in <graph break in forward_and_backward_pass> | |
pred = mod(*cloned_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "/data/users/ezyang/benchmark/torchbenchmark/util/framework/huggingface/model_factory.py", line 41, in forward | |
return self.model(input_ids=input_ids, decoder_input_ids=decoder_input_ids) | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/t5/modeling_t5.py", line 1537, in forward | |
@add_start_docstrings_to_model_forward(T5_INPUTS_DOCSTRING) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 893, in forward | |
return compiled_f( | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 880, in new_func | |
compiled_fn = create_aot_dispatcher_function( | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 600, in create_aot_dispatcher_function | |
aot_dispatch_autograd(flat_fn, fake_flat_tensor_args, aot_config) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 459, in aot_dispatch_autograd | |
compiled_fw_func = aot_config.fw_compiler(fw_module, deduped_flat_args) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 351, in fw_compiler | |
return compile_fx_inner( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/debug_utils.py", line 466, in debug_wrapper | |
compiled_fn = compiler_fn(gm, example_inputs, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/debug.py", line 177, in inner | |
return fn(*args, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/contextlib.py", line 79, in inner | |
return func(*args, **kwds) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 122, in compile_fx_inner | |
graph.run(*example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 129, in run | |
return super().run(*args) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 130, in run | |
self.env[node] = self.run_node(node) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 299, in run_node | |
result = super().run_node(n) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 171, in run_node | |
return getattr(self, n.op)(n.target, args, kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 242, in call_function | |
raise LoweringException(e, target, args, kwargs) from e | |
torch._inductor.exc.LoweringException: RuntimeError: Overloaded torch operator invoked from Python failed to many any schema: | |
aten::arange() expected at most 5 argument(s) but received 7 argument(s). Declaration: aten::arange(Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() expected at most 6 argument(s) but received 7 argument(s). Declaration: aten::arange.start(Scalar start, Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() Expected a value of type 'number' for argument 'end' but instead found type 'Symbol'. | |
Position: 1 | |
Value: s1 | |
Declaration: aten::arange.start_step(Scalar start, Scalar end, Scalar step=1, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
Cast error details: Cannot cast s1 to number | |
aten::arange() expected at most 4 argument(s) but received 7 argument(s). Declaration: aten::arange.start_out(Scalar start, Scalar end, Scalar step=1, *, Tensor(a!) out) -> Tensor(a!) | |
aten::arange() expected at most 2 argument(s) but received 7 argument(s). Declaration: aten::arange.out(Scalar end, *, Tensor(a!) out) -> Tensor(a!) | |
target: aten.arange.default | |
args[0]: s1 | |
kwargs: {'dtype': torch.int64, 'device': device(type='cuda', index=0), 'pin_memory': False} | |
While executing %arange : [#users=2] = call_function[target=torch.ops.aten.arange.default](args = (%sym_size_2,), kwargs = {dtype: torch.int64, device: cuda:0, pin_memory: False}) | |
Original traceback: | |
Module stack: {'self_encoder': <class 'transformers.models.t5.modeling_t5.T5Stack'>, 'self_encoder_block_0': <class 'transformers.models.t5.modeling_t5.T5Block'>, 'sub0_0': <class 'transformers.models.t5.modeling_t5.T5LayerSelfAttention'>, 'self_encoder_block_0_layer_0_SelfAttention': <class 'transformers.models.t5.modeling_t5.T5Attention'>} | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/t5/modeling_t5.py", line 423, in compute_bias | |
context_position = torch.arange(query_length, dtype=torch.long, device=device)[:, None] | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/t5/modeling_t5.py", line 519, in forward | |
position_bias = self.compute_bias(real_seq_length, key_length, device=scores.device) | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/t5/modeling_t5.py", line 570, in forward | |
attention_output = self.SelfAttention( | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/t5/modeling_t5.py", line 664, in forward | |
self_attention_outputs = self.layer[0]( | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/t5/modeling_t5.py", line 1033, in forward | |
layer_outputs = layer_module( | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/t5/modeling_t5.py", line 1601, in forward | |
encoder_outputs = self.encoder( | |
TorchDynamo optimized model failed to run because of following error | |
cuda train hf_T5 FAIL | |
Running torchbench.py hf_T5_base... | |
WARNING:common:fp64 golden ref were not generated for hf_T5_base | |
ERROR:common:RuntimeError: Overloaded torch operator invoked from Python failed to many any schema: | |
aten::arange() expected at most 5 argument(s) but received 7 argument(s). Declaration: aten::arange(Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() expected at most 6 argument(s) but received 7 argument(s). Declaration: aten::arange.start(Scalar start, Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() Expected a value of type 'number' for argument 'end' but instead found type 'Symbol'. | |
Position: 1 | |
Value: s1 | |
Declaration: aten::arange.start_step(Scalar start, Scalar end, Scalar step=1, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
Cast error details: Cannot cast s1 to number | |
aten::arange() expected at most 4 argument(s) but received 7 argument(s). Declaration: aten::arange.start_out(Scalar start, Scalar end, Scalar step=1, *, Tensor(a!) out) -> Tensor(a!) | |
aten::arange() expected at most 2 argument(s) but received 7 argument(s). Declaration: aten::arange.out(Scalar end, *, Tensor(a!) out) -> Tensor(a!) | |
target: aten.arange.default | |
args[0]: s1 | |
kwargs: {'dtype': torch.int64, 'device': device(type='cuda', index=0), 'pin_memory': False} | |
While executing %arange : [#users=2] = call_function[target=torch.ops.aten.arange.default](args = (%sym_size_2,), kwargs = {dtype: torch.int64, device: cuda:0, pin_memory: False}) | |
Original traceback: | |
Module stack: {'self_encoder': <class 'transformers.models.t5.modeling_t5.T5Stack'>, 'self_encoder_block_0': <class 'transformers.models.t5.modeling_t5.T5Block'>, 'sub0_0': <class 'transformers.models.t5.modeling_t5.T5LayerSelfAttention'>, 'self_encoder_block_0_layer_0_SelfAttention': <class 'transformers.models.t5.modeling_t5.T5Attention'>} | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/t5/modeling_t5.py", line 423, in compute_bias | |
context_position = torch.arange(query_length, dtype=torch.long, device=device)[:, None] | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/t5/modeling_t5.py", line 519, in forward | |
position_bias = self.compute_bias(real_seq_length, key_length, device=scores.device) | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/t5/modeling_t5.py", line 570, in forward | |
attention_output = self.SelfAttention( | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/t5/modeling_t5.py", line 664, in forward | |
self_attention_outputs = self.layer[0]( | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/t5/modeling_t5.py", line 1033, in forward | |
layer_outputs = layer_module( | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/t5/modeling_t5.py", line 1601, in forward | |
encoder_outputs = self.encoder( | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 239, in call_function | |
out = lowerings[target](*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/lowering.py", line 204, in wrapped | |
return decomp_fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/lowering.py", line 1194, in arange | |
return fallback_arange( | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/lowering.py", line 893, in handler | |
result = ir.FallbackKernel.create(kernel, *args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/ir.py", line 2890, in create | |
example_output = kernel( | |
File "/data/users/ezyang/pytorch-tmp/torch/_ops.py", line 446, in __call__ | |
return self._op(*args, **kwargs or {}) | |
RuntimeError: Overloaded torch operator invoked from Python failed to many any schema: | |
aten::arange() expected at most 5 argument(s) but received 7 argument(s). Declaration: aten::arange(Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() expected at most 6 argument(s) but received 7 argument(s). Declaration: aten::arange.start(Scalar start, Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() Expected a value of type 'number' for argument 'end' but instead found type 'Symbol'. | |
Position: 1 | |
Value: s1 | |
Declaration: aten::arange.start_step(Scalar start, Scalar end, Scalar step=1, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
Cast error details: Cannot cast s1 to number | |
aten::arange() expected at most 4 argument(s) but received 7 argument(s). Declaration: aten::arange.start_out(Scalar start, Scalar end, Scalar step=1, *, Tensor(a!) out) -> Tensor(a!) | |
aten::arange() expected at most 2 argument(s) but received 7 argument(s). Declaration: aten::arange.out(Scalar end, *, Tensor(a!) out) -> Tensor(a!) | |
The above exception was the direct cause of the following exception: | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1041, in check_accuracy | |
new_result = optimized_model_iter_fn(model, example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 945, in run_n_iterations | |
self.model_iter_fn(mod, inputs, collect_outputs=False) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 332, in forward_and_backward_pass | |
cloned_inputs = clone_inputs(inputs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 335, in <graph break in forward_and_backward_pass> | |
pred = mod(*cloned_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "/data/users/ezyang/benchmark/torchbenchmark/util/framework/huggingface/model_factory.py", line 41, in forward | |
return self.model(input_ids=input_ids, decoder_input_ids=decoder_input_ids) | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/t5/modeling_t5.py", line 1537, in forward | |
@add_start_docstrings_to_model_forward(T5_INPUTS_DOCSTRING) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 893, in forward | |
return compiled_f( | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 880, in new_func | |
compiled_fn = create_aot_dispatcher_function( | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 600, in create_aot_dispatcher_function | |
aot_dispatch_autograd(flat_fn, fake_flat_tensor_args, aot_config) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 459, in aot_dispatch_autograd | |
compiled_fw_func = aot_config.fw_compiler(fw_module, deduped_flat_args) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 351, in fw_compiler | |
return compile_fx_inner( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/debug_utils.py", line 466, in debug_wrapper | |
compiled_fn = compiler_fn(gm, example_inputs, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/debug.py", line 177, in inner | |
return fn(*args, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/contextlib.py", line 79, in inner | |
return func(*args, **kwds) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 122, in compile_fx_inner | |
graph.run(*example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 129, in run | |
return super().run(*args) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 130, in run | |
self.env[node] = self.run_node(node) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 299, in run_node | |
result = super().run_node(n) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 171, in run_node | |
return getattr(self, n.op)(n.target, args, kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 242, in call_function | |
raise LoweringException(e, target, args, kwargs) from e | |
torch._inductor.exc.LoweringException: RuntimeError: Overloaded torch operator invoked from Python failed to many any schema: | |
aten::arange() expected at most 5 argument(s) but received 7 argument(s). Declaration: aten::arange(Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() expected at most 6 argument(s) but received 7 argument(s). Declaration: aten::arange.start(Scalar start, Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() Expected a value of type 'number' for argument 'end' but instead found type 'Symbol'. | |
Position: 1 | |
Value: s1 | |
Declaration: aten::arange.start_step(Scalar start, Scalar end, Scalar step=1, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
Cast error details: Cannot cast s1 to number | |
aten::arange() expected at most 4 argument(s) but received 7 argument(s). Declaration: aten::arange.start_out(Scalar start, Scalar end, Scalar step=1, *, Tensor(a!) out) -> Tensor(a!) | |
aten::arange() expected at most 2 argument(s) but received 7 argument(s). Declaration: aten::arange.out(Scalar end, *, Tensor(a!) out) -> Tensor(a!) | |
target: aten.arange.default | |
args[0]: s1 | |
kwargs: {'dtype': torch.int64, 'device': device(type='cuda', index=0), 'pin_memory': False} | |
While executing %arange : [#users=2] = call_function[target=torch.ops.aten.arange.default](args = (%sym_size_2,), kwargs = {dtype: torch.int64, device: cuda:0, pin_memory: False}) | |
Original traceback: | |
Module stack: {'self_encoder': <class 'transformers.models.t5.modeling_t5.T5Stack'>, 'self_encoder_block_0': <class 'transformers.models.t5.modeling_t5.T5Block'>, 'sub0_0': <class 'transformers.models.t5.modeling_t5.T5LayerSelfAttention'>, 'self_encoder_block_0_layer_0_SelfAttention': <class 'transformers.models.t5.modeling_t5.T5Attention'>} | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/t5/modeling_t5.py", line 423, in compute_bias | |
context_position = torch.arange(query_length, dtype=torch.long, device=device)[:, None] | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/t5/modeling_t5.py", line 519, in forward | |
position_bias = self.compute_bias(real_seq_length, key_length, device=scores.device) | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/t5/modeling_t5.py", line 570, in forward | |
attention_output = self.SelfAttention( | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/t5/modeling_t5.py", line 664, in forward | |
self_attention_outputs = self.layer[0]( | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/t5/modeling_t5.py", line 1033, in forward | |
layer_outputs = layer_module( | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/t5/modeling_t5.py", line 1601, in forward | |
encoder_outputs = self.encoder( | |
TorchDynamo optimized model failed to run because of following error | |
cuda train hf_T5_base FAIL | |
Running torchbench.py hf_T5_large... | |
WARNING:common:fp64 golden ref were not generated for hf_T5_large | |
ERROR:common:RuntimeError: Overloaded torch operator invoked from Python failed to many any schema: | |
aten::arange() expected at most 5 argument(s) but received 7 argument(s). Declaration: aten::arange(Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() expected at most 6 argument(s) but received 7 argument(s). Declaration: aten::arange.start(Scalar start, Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() Expected a value of type 'number' for argument 'end' but instead found type 'Symbol'. | |
Position: 1 | |
Value: s1 | |
Declaration: aten::arange.start_step(Scalar start, Scalar end, Scalar step=1, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
Cast error details: Cannot cast s1 to number | |
aten::arange() expected at most 4 argument(s) but received 7 argument(s). Declaration: aten::arange.start_out(Scalar start, Scalar end, Scalar step=1, *, Tensor(a!) out) -> Tensor(a!) | |
aten::arange() expected at most 2 argument(s) but received 7 argument(s). Declaration: aten::arange.out(Scalar end, *, Tensor(a!) out) -> Tensor(a!) | |
target: aten.arange.default | |
args[0]: s1 | |
kwargs: {'dtype': torch.int64, 'device': device(type='cuda', index=0), 'pin_memory': False} | |
While executing %arange : [#users=2] = call_function[target=torch.ops.aten.arange.default](args = (%sym_size_2,), kwargs = {dtype: torch.int64, device: cuda:0, pin_memory: False}) | |
Original traceback: | |
Module stack: {'self_encoder': <class 'transformers.models.t5.modeling_t5.T5Stack'>, 'self_encoder_block_0': <class 'transformers.models.t5.modeling_t5.T5Block'>, 'sub0_0': <class 'transformers.models.t5.modeling_t5.T5LayerSelfAttention'>, 'self_encoder_block_0_layer_0_SelfAttention': <class 'transformers.models.t5.modeling_t5.T5Attention'>} | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/t5/modeling_t5.py", line 423, in compute_bias | |
context_position = torch.arange(query_length, dtype=torch.long, device=device)[:, None] | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/t5/modeling_t5.py", line 519, in forward | |
position_bias = self.compute_bias(real_seq_length, key_length, device=scores.device) | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/t5/modeling_t5.py", line 570, in forward | |
attention_output = self.SelfAttention( | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/t5/modeling_t5.py", line 664, in forward | |
self_attention_outputs = self.layer[0]( | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/t5/modeling_t5.py", line 1033, in forward | |
layer_outputs = layer_module( | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/t5/modeling_t5.py", line 1601, in forward | |
encoder_outputs = self.encoder( | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 239, in call_function | |
out = lowerings[target](*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/lowering.py", line 204, in wrapped | |
return decomp_fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/lowering.py", line 1194, in arange | |
return fallback_arange( | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/lowering.py", line 893, in handler | |
result = ir.FallbackKernel.create(kernel, *args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/ir.py", line 2890, in create | |
example_output = kernel( | |
File "/data/users/ezyang/pytorch-tmp/torch/_ops.py", line 446, in __call__ | |
return self._op(*args, **kwargs or {}) | |
RuntimeError: Overloaded torch operator invoked from Python failed to many any schema: | |
aten::arange() expected at most 5 argument(s) but received 7 argument(s). Declaration: aten::arange(Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() expected at most 6 argument(s) but received 7 argument(s). Declaration: aten::arange.start(Scalar start, Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() Expected a value of type 'number' for argument 'end' but instead found type 'Symbol'. | |
Position: 1 | |
Value: s1 | |
Declaration: aten::arange.start_step(Scalar start, Scalar end, Scalar step=1, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
Cast error details: Cannot cast s1 to number | |
aten::arange() expected at most 4 argument(s) but received 7 argument(s). Declaration: aten::arange.start_out(Scalar start, Scalar end, Scalar step=1, *, Tensor(a!) out) -> Tensor(a!) | |
aten::arange() expected at most 2 argument(s) but received 7 argument(s). Declaration: aten::arange.out(Scalar end, *, Tensor(a!) out) -> Tensor(a!) | |
The above exception was the direct cause of the following exception: | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1041, in check_accuracy | |
new_result = optimized_model_iter_fn(model, example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 945, in run_n_iterations | |
self.model_iter_fn(mod, inputs, collect_outputs=False) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 332, in forward_and_backward_pass | |
cloned_inputs = clone_inputs(inputs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 335, in <graph break in forward_and_backward_pass> | |
pred = mod(*cloned_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "/data/users/ezyang/benchmark/torchbenchmark/util/framework/huggingface/model_factory.py", line 41, in forward | |
return self.model(input_ids=input_ids, decoder_input_ids=decoder_input_ids) | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/t5/modeling_t5.py", line 1537, in forward | |
@add_start_docstrings_to_model_forward(T5_INPUTS_DOCSTRING) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 893, in forward | |
return compiled_f( | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 880, in new_func | |
compiled_fn = create_aot_dispatcher_function( | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 600, in create_aot_dispatcher_function | |
aot_dispatch_autograd(flat_fn, fake_flat_tensor_args, aot_config) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 459, in aot_dispatch_autograd | |
compiled_fw_func = aot_config.fw_compiler(fw_module, deduped_flat_args) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 351, in fw_compiler | |
return compile_fx_inner( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/debug_utils.py", line 466, in debug_wrapper | |
compiled_fn = compiler_fn(gm, example_inputs, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/debug.py", line 177, in inner | |
return fn(*args, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/contextlib.py", line 79, in inner | |
return func(*args, **kwds) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 122, in compile_fx_inner | |
graph.run(*example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 129, in run | |
return super().run(*args) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 130, in run | |
self.env[node] = self.run_node(node) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 299, in run_node | |
result = super().run_node(n) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 171, in run_node | |
return getattr(self, n.op)(n.target, args, kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 242, in call_function | |
raise LoweringException(e, target, args, kwargs) from e | |
torch._inductor.exc.LoweringException: RuntimeError: Overloaded torch operator invoked from Python failed to many any schema: | |
aten::arange() expected at most 5 argument(s) but received 7 argument(s). Declaration: aten::arange(Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() expected at most 6 argument(s) but received 7 argument(s). Declaration: aten::arange.start(Scalar start, Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() Expected a value of type 'number' for argument 'end' but instead found type 'Symbol'. | |
Position: 1 | |
Value: s1 | |
Declaration: aten::arange.start_step(Scalar start, Scalar end, Scalar step=1, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
Cast error details: Cannot cast s1 to number | |
aten::arange() expected at most 4 argument(s) but received 7 argument(s). Declaration: aten::arange.start_out(Scalar start, Scalar end, Scalar step=1, *, Tensor(a!) out) -> Tensor(a!) | |
aten::arange() expected at most 2 argument(s) but received 7 argument(s). Declaration: aten::arange.out(Scalar end, *, Tensor(a!) out) -> Tensor(a!) | |
target: aten.arange.default | |
args[0]: s1 | |
kwargs: {'dtype': torch.int64, 'device': device(type='cuda', index=0), 'pin_memory': False} | |
While executing %arange : [#users=2] = call_function[target=torch.ops.aten.arange.default](args = (%sym_size_2,), kwargs = {dtype: torch.int64, device: cuda:0, pin_memory: False}) | |
Original traceback: | |
Module stack: {'self_encoder': <class 'transformers.models.t5.modeling_t5.T5Stack'>, 'self_encoder_block_0': <class 'transformers.models.t5.modeling_t5.T5Block'>, 'sub0_0': <class 'transformers.models.t5.modeling_t5.T5LayerSelfAttention'>, 'self_encoder_block_0_layer_0_SelfAttention': <class 'transformers.models.t5.modeling_t5.T5Attention'>} | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/t5/modeling_t5.py", line 423, in compute_bias | |
context_position = torch.arange(query_length, dtype=torch.long, device=device)[:, None] | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/t5/modeling_t5.py", line 519, in forward | |
position_bias = self.compute_bias(real_seq_length, key_length, device=scores.device) | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/t5/modeling_t5.py", line 570, in forward | |
attention_output = self.SelfAttention( | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/t5/modeling_t5.py", line 664, in forward | |
self_attention_outputs = self.layer[0]( | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/t5/modeling_t5.py", line 1033, in forward | |
layer_outputs = layer_module( | |
| File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/site-packages/transformers/models/t5/modeling_t5.py", line 1601, in forward | |
encoder_outputs = self.encoder( | |
TorchDynamo optimized model failed to run because of following error | |
cuda train hf_T5_large FAIL | |
Running torchbench.py lennard_jones... | |
cuda train lennard_jones PASS | |
Running torchbench.py maml_omniglot... | |
ERROR:common:'int' object has no attribute 'size' | |
While executing %sym_size : [#users=1] = placeholder[target=sym_size] | |
Original traceback: | |
Module stack: {'self_12': <class 'torch.nn.modules.flatten.Flatten'>} | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/container.py", line 204, in forward | |
input = module(input) | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1041, in check_accuracy | |
new_result = optimized_model_iter_fn(model, example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 945, in run_n_iterations | |
self.model_iter_fn(mod, inputs, collect_outputs=False) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 332, in forward_and_backward_pass | |
cloned_inputs = clone_inputs(inputs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 337, in <graph break in forward_and_backward_pass> | |
self.grad_scaler.scale(loss).backward() | |
File "/data/users/ezyang/pytorch-tmp/torch/_tensor.py", line 488, in backward | |
torch.autograd.backward( | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/__init__.py", line 197, in backward | |
Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/function.py", line 270, in apply | |
return user_fn(self, *args) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 499, in backward | |
CompiledFunction.compiled_bw = aot_config.bw_compiler( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/optimizations/backends.py", line 555, in _wrapped_bw_compiler | |
return disable(bw_compiler(*args, **kwargs)) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 362, in bw_compiler | |
return compile_fx_inner( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/debug_utils.py", line 466, in debug_wrapper | |
compiled_fn = compiler_fn(gm, example_inputs, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/debug.py", line 177, in inner | |
return fn(*args, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/contextlib.py", line 79, in inner | |
return func(*args, **kwds) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 122, in compile_fx_inner | |
graph.run(*example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 129, in run | |
return super().run(*args) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 130, in run | |
self.env[node] = self.run_node(node) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 299, in run_node | |
result = super().run_node(n) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 171, in run_node | |
return getattr(self, n.op)(n.target, args, kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 199, in placeholder | |
sizes, strides = self.static_sizes_strides(example) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 53, in static_sizes_strides | |
size = [sympy.Integer(i) for i in ex.size()] | |
AttributeError: 'int' object has no attribute 'size' | |
While executing %sym_size : [#users=1] = placeholder[target=sym_size] | |
Original traceback: | |
Module stack: {'self_12': <class 'torch.nn.modules.flatten.Flatten'>} | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/container.py", line 204, in forward | |
input = module(input) | |
TorchDynamo optimized model failed to run because of following error | |
cuda train maml_omniglot FAIL | |
Running torchbench.py mnasnet1_0... | |
ERROR:common:AssertionError: | |
target: aten.div.Scalar | |
args[0]: TensorBox( | |
ReinterpretView( | |
StorageBox( | |
MatrixMultiply( | |
name=buf0, | |
layout=FixedLayout('cuda', torch.float32, size=[s0, 1280], stride=[1280, 1]), | |
inputs=[InputBuffer(name='orig_tangents_1', layout=FixedLayout('cuda', torch.float32, size=[s0, s1], stride=[s1, 1])), InputBuffer(name='permute_1', layout=FixedLayout('cuda', torch.float32, size=[1000, 1280], stride=[1280, 1]))], | |
constant_args=(), | |
kwargs={}, | |
output_view=None, | |
origins={mm} | |
) | |
), | |
FixedLayout('cuda', torch.float32, size=[2, 1280, 7, 7], stride=[1280, 1, 0, 0]), | |
origins={expand} | |
) | |
) | |
args[1]: 49 | |
While executing %div : [#users=1] = call_function[target=torch.ops.aten.div.Scalar](args = (%expand, %mul_156), kwargs = {}) | |
Original traceback: | |
Module stack: {} | |
File "/data/users/ezyang/vision/torchvision/models/mnasnet.py", line 161, in forward | |
x = x.mean([2, 3]) | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 239, in call_function | |
out = lowerings[target](*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/lowering.py", line 204, in wrapped | |
return decomp_fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/lowering.py", line 3182, in div | |
dtype = get_promoted_dtype( | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/lowering.py", line 138, in get_promoted_dtype | |
inps = [construct_input(arg) for arg in args] | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/lowering.py", line 138, in <listcomp> | |
inps = [construct_input(arg) for arg in args] | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/lowering.py", line 133, in construct_input | |
assert hasattr(inp, "get_dtype") | |
AssertionError | |
The above exception was the direct cause of the following exception: | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1041, in check_accuracy | |
new_result = optimized_model_iter_fn(model, example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 945, in run_n_iterations | |
self.model_iter_fn(mod, inputs, collect_outputs=False) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 332, in forward_and_backward_pass | |
cloned_inputs = clone_inputs(inputs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 337, in <graph break in forward_and_backward_pass> | |
self.grad_scaler.scale(loss).backward() | |
File "/data/users/ezyang/pytorch-tmp/torch/_tensor.py", line 488, in backward | |
torch.autograd.backward( | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/__init__.py", line 197, in backward | |
Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/function.py", line 270, in apply | |
return user_fn(self, *args) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 499, in backward | |
CompiledFunction.compiled_bw = aot_config.bw_compiler( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/optimizations/backends.py", line 555, in _wrapped_bw_compiler | |
return disable(bw_compiler(*args, **kwargs)) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 362, in bw_compiler | |
return compile_fx_inner( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/debug_utils.py", line 466, in debug_wrapper | |
compiled_fn = compiler_fn(gm, example_inputs, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/debug.py", line 177, in inner | |
return fn(*args, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/contextlib.py", line 79, in inner | |
return func(*args, **kwds) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 122, in compile_fx_inner | |
graph.run(*example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 129, in run | |
return super().run(*args) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 130, in run | |
self.env[node] = self.run_node(node) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 299, in run_node | |
result = super().run_node(n) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 171, in run_node | |
return getattr(self, n.op)(n.target, args, kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 242, in call_function | |
raise LoweringException(e, target, args, kwargs) from e | |
torch._inductor.exc.LoweringException: AssertionError: | |
target: aten.div.Scalar | |
args[0]: TensorBox( | |
ReinterpretView( | |
StorageBox( | |
MatrixMultiply( | |
name=buf0, | |
layout=FixedLayout('cuda', torch.float32, size=[s0, 1280], stride=[1280, 1]), | |
inputs=[InputBuffer(name='orig_tangents_1', layout=FixedLayout('cuda', torch.float32, size=[s0, s1], stride=[s1, 1])), InputBuffer(name='permute_1', layout=FixedLayout('cuda', torch.float32, size=[1000, 1280], stride=[1280, 1]))], | |
constant_args=(), | |
kwargs={}, | |
output_view=None, | |
origins={mm} | |
) | |
), | |
FixedLayout('cuda', torch.float32, size=[2, 1280, 7, 7], stride=[1280, 1, 0, 0]), | |
origins={expand} | |
) | |
) | |
args[1]: 49 | |
While executing %div : [#users=1] = call_function[target=torch.ops.aten.div.Scalar](args = (%expand, %mul_156), kwargs = {}) | |
Original traceback: | |
Module stack: {} | |
File "/data/users/ezyang/vision/torchvision/models/mnasnet.py", line 161, in forward | |
x = x.mean([2, 3]) | |
TorchDynamo optimized model failed to run because of following error | |
cuda train mnasnet1_0 FAIL | |
Running torchbench.py mobilenet_v2... | |
ERROR:common:'int' object has no attribute 'size' | |
While executing %sym_size : [#users=2] = placeholder[target=sym_size] | |
Original traceback: | |
Module stack: {} | |
File "/data/users/ezyang/vision/torchvision/models/mobilenetv2.py", line 169, in _forward_impl | |
x = torch.flatten(x, 1) | |
| File "/data/users/ezyang/vision/torchvision/models/mobilenetv2.py", line 174, in forward | |
return self._forward_impl(x) | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1041, in check_accuracy | |
new_result = optimized_model_iter_fn(model, example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 945, in run_n_iterations | |
self.model_iter_fn(mod, inputs, collect_outputs=False) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 332, in forward_and_backward_pass | |
cloned_inputs = clone_inputs(inputs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 337, in <graph break in forward_and_backward_pass> | |
self.grad_scaler.scale(loss).backward() | |
File "/data/users/ezyang/pytorch-tmp/torch/_tensor.py", line 488, in backward | |
torch.autograd.backward( | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/__init__.py", line 197, in backward | |
Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/function.py", line 270, in apply | |
return user_fn(self, *args) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 499, in backward | |
CompiledFunction.compiled_bw = aot_config.bw_compiler( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/optimizations/backends.py", line 555, in _wrapped_bw_compiler | |
return disable(bw_compiler(*args, **kwargs)) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 362, in bw_compiler | |
return compile_fx_inner( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/debug_utils.py", line 466, in debug_wrapper | |
compiled_fn = compiler_fn(gm, example_inputs, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/debug.py", line 177, in inner | |
return fn(*args, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/contextlib.py", line 79, in inner | |
return func(*args, **kwds) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 122, in compile_fx_inner | |
graph.run(*example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 129, in run | |
return super().run(*args) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 130, in run | |
self.env[node] = self.run_node(node) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 299, in run_node | |
result = super().run_node(n) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 171, in run_node | |
return getattr(self, n.op)(n.target, args, kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 199, in placeholder | |
sizes, strides = self.static_sizes_strides(example) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 53, in static_sizes_strides | |
size = [sympy.Integer(i) for i in ex.size()] | |
AttributeError: 'int' object has no attribute 'size' | |
While executing %sym_size : [#users=2] = placeholder[target=sym_size] | |
Original traceback: | |
Module stack: {} | |
File "/data/users/ezyang/vision/torchvision/models/mobilenetv2.py", line 169, in _forward_impl | |
x = torch.flatten(x, 1) | |
| File "/data/users/ezyang/vision/torchvision/models/mobilenetv2.py", line 174, in forward | |
return self._forward_impl(x) | |
TorchDynamo optimized model failed to run because of following error | |
cuda train mobilenet_v2 FAIL | |
Running torchbench.py mobilenet_v2_quantized_qat... | |
WARNING:common:fp64 golden ref were not generated for mobilenet_v2_quantized_qat | |
[2022-10-30 21:28:29,841] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:29,848] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:29,862] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:29,866] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:29,871] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:29,882] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:29,886] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:29,890] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:29,902] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:29,906] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:29,918] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:29,922] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:29,926] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:29,939] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:29,943] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:29,947] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:29,959] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:29,963] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:29,975] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:29,979] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:29,983] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:29,995] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:29,999] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,003] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,015] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,017] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,021] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,033] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,037] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,041] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,054] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,058] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,062] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,075] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,079] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,092] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,096] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,100] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,112] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,116] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,120] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,133] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,135] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,138] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,150] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,154] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,158] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,169] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,173] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,177] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,189] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,192] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,195] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,207] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,211] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,215] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,227] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,231] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,234] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,247] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,251] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,263] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,267] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,271] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,282] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,286] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,290] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,301] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,303] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,307] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,318] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,322] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,326] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,337] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,341] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,345] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,356] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,358] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,362] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,373] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,377] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,381] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,393] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,397] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,400] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,412] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,414] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,418] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,429] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,434] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,437] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,449] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,453] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,457] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,469] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,472] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,485] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,488] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,492] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,504] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,508] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,512] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,523] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,525] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,529] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,541] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,544] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,548] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,559] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,563] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,567] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,578] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,581] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,584] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,596] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,599] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,603] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,616] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,620] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,624] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,636] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,640] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,653] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,657] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,661] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,672] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,675] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,680] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,691] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,693] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,697] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,708] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,712] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,716] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,727] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,732] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,735] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,747] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,749] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,752] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,764] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,768] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,772] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,784] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,788] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:30,791] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:31,307] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:31,311] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:31,323] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:31,327] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:31,331] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:31,333] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:31,334] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:31,335] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:28:31,336] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
ERROR:common:name 's0' is not defined | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1041, in check_accuracy | |
new_result = optimized_model_iter_fn(model, example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 945, in run_n_iterations | |
self.model_iter_fn(mod, inputs, collect_outputs=False) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 332, in forward_and_backward_pass | |
cloned_inputs = clone_inputs(inputs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 335, in <graph break in forward_and_backward_pass> | |
pred = mod(*cloned_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/graph_module.py", line 660, in call_wrapped | |
return self._wrapped_call(self, *args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/graph_module.py", line 279, in __call__ | |
raise e | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/graph_module.py", line 269, in __call__ | |
return super(self.cls, obj).__call__(*args, **kwargs) # type: ignore[misc] | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "<eval_with_key>.5", line 4, in forward | |
def forward(self, x : torch.Tensor) -> torch.Tensor: | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 893, in forward | |
return compiled_f( | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 885, in new_func | |
return compiled_fn(args) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 271, in g | |
return f(*args) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 513, in compiled_function | |
return CompiledFunction.apply(*remove_dupe_args(args)) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 471, in forward | |
fw_outs = call_func_with_args( | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 296, in call_func_with_args | |
out = normalize_as_list(f(args)) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 186, in run | |
return model(new_inputs) | |
File "/tmp/torchinductor_ezyang/cl/cclnalhvewt4b3vliempyodfzqlczg4crwi2xomahb6n7lu3hu72.py", line 4488, in call | |
return (buf1283, orig_primals_1, orig_primals_2, orig_primals_4, orig_primals_5, orig_primals_7, orig_primals_8, orig_primals_10, orig_primals_11, orig_primals_13, orig_primals_14, orig_primals_16, orig_primals_17, orig_primals_19, orig_primals_20, orig_primals_22, orig_primals_23, orig_primals_25, orig_primals_26, orig_primals_28, orig_primals_29, orig_primals_31, orig_primals_32, orig_primals_34, orig_primals_35, orig_primals_37, orig_primals_38, orig_primals_40, orig_primals_41, orig_primals_43, orig_primals_44, orig_primals_46, orig_primals_47, orig_primals_49, orig_primals_50, orig_primals_52, orig_primals_53, orig_primals_55, orig_primals_56, orig_primals_58, orig_primals_59, orig_primals_61, orig_primals_62, orig_primals_64, orig_primals_65, orig_primals_67, orig_primals_68, orig_primals_70, orig_primals_71, orig_primals_73, orig_primals_74, orig_primals_76, orig_primals_77, orig_primals_79, orig_primals_80, orig_primals_82, orig_primals_83, orig_primals_85, orig_primals_86, orig_primals_88, orig_primals_89, orig_primals_91, orig_primals_92, orig_primals_94, orig_primals_95, orig_primals_97, orig_primals_98, orig_primals_100, orig_primals_101, orig_primals_103, orig_primals_104, orig_primals_106, orig_primals_107, orig_primals_109, orig_primals_110, orig_primals_112, orig_primals_113, orig_primals_115, orig_primals_116, orig_primals_118, orig_primals_119, orig_primals_121, orig_primals_122, orig_primals_124, orig_primals_125, orig_primals_127, orig_primals_128, orig_primals_130, orig_primals_131, orig_primals_133, orig_primals_134, orig_primals_136, orig_primals_137, orig_primals_139, orig_primals_140, orig_primals_142, orig_primals_143, orig_primals_145, orig_primals_146, orig_primals_148, orig_primals_149, orig_primals_151, orig_primals_152, orig_primals_154, orig_primals_155, orig_primals_166, orig_primals_167, orig_primals_183, orig_primals_184, orig_primals_200, orig_primals_201, orig_primals_217, orig_primals_218, orig_primals_234, orig_primals_235, orig_primals_251, orig_primals_252, orig_primals_268, orig_primals_269, orig_primals_285, orig_primals_286, orig_primals_302, orig_primals_303, orig_primals_326, orig_primals_327, orig_primals_343, orig_primals_344, orig_primals_360, orig_primals_361, orig_primals_377, orig_primals_378, orig_primals_394, orig_primals_395, orig_primals_411, orig_primals_412, orig_primals_435, orig_primals_436, orig_primals_452, orig_primals_453, orig_primals_469, orig_primals_470, orig_primals_493, orig_primals_494, orig_primals_510, orig_primals_511, orig_primals_527, orig_primals_528, orig_primals_544, orig_primals_545, orig_primals_561, orig_primals_562, orig_primals_578, orig_primals_579, orig_primals_602, orig_primals_603, orig_primals_619, orig_primals_620, orig_primals_636, orig_primals_637, orig_primals_660, orig_primals_661, orig_primals_677, orig_primals_678, orig_primals_694, orig_primals_695, orig_primals_718, orig_primals_719, orig_primals_735, orig_primals_736, orig_primals_752, orig_primals_753, orig_primals_769, orig_primals_770, orig_primals_786, orig_primals_787, orig_primals_803, orig_primals_804, orig_primals_827, orig_primals_828, orig_primals_844, orig_primals_845, orig_primals_861, orig_primals_862, orig_primals_885, orig_primals_886, orig_primals_902, orig_primals_903, orig_primals_919, orig_primals_920, orig_primals_936, orig_primals_937, orig_primals_953, orig_primals_954, orig_primals_970, orig_primals_971, orig_primals_994, orig_primals_995, orig_primals_1011, orig_primals_1012, orig_primals_1028, orig_primals_1029, orig_primals_1052, orig_primals_1053, orig_primals_1069, orig_primals_1070, orig_primals_1086, orig_primals_1087, orig_primals_1103, orig_primals_1104, buf1, buf9, buf10, buf1289, buf19, buf26, buf27, buf34, buf35, buf1290, buf44, buf51, buf52, buf59, buf60, buf1291, buf68, buf69, buf76, buf77, buf1292, buf86, buf93, buf94, buf101, buf102, buf1293, buf111, buf118, buf119, buf126, buf127, buf1294, buf135, buf136, buf143, buf144, buf1295, buf153, buf160, buf161, buf168, buf169, buf1296, buf178, buf185, buf186, buf193, buf194, buf1297, buf203, buf210, buf211, buf218, buf219, buf1298, buf228, buf235, buf236, buf243, buf244, buf1299, buf253, buf260, buf261, buf268, buf269, buf1300, buf277, buf278, buf285, buf286, buf1301, buf295, buf302, buf303, buf310, buf311, buf1302, buf320, buf327, buf328, buf335, buf336, buf1303, buf345, buf352, buf353, buf360, buf361, buf1304, buf370, buf377, buf378, buf385, buf386, buf1305, buf395, buf402, buf403, buf410, buf411, buf1306, buf420, buf427, buf428, buf435, buf436, buf1307, buf445, buf452, buf453, buf460, buf461, buf1308, buf470, buf477, buf478, buf485, buf486, buf1309, buf494, buf495, buf502, buf503, buf1310, buf512, buf519, buf520, buf527, buf528, buf1311, buf537, buf544, buf545, buf552, buf553, buf1312, buf562, buf569, buf570, buf577, buf578, buf1313, buf587, buf594, buf595, buf602, buf603, buf1314, buf612, buf619, buf620, buf627, buf628, buf1315, buf637, buf644, buf645, buf652, buf653, buf1316, buf662, buf669, buf670, buf677, buf678, buf1317, buf687, buf694, buf695, buf702, buf703, buf1318, buf712, buf719, buf720, buf727, buf728, buf1319, buf737, buf744, buf745, buf752, buf753, buf1320, buf762, buf769, buf770, buf777, buf778, buf1321, buf786, buf787, buf794, buf795, buf1322, buf804, buf811, buf812, buf819, buf820, buf1323, buf829, buf836, buf837, buf844, buf845, buf1324, buf854, buf861, buf862, buf869, buf870, buf1325, buf879, buf886, buf887, buf894, buf895, buf1326, buf904, buf911, buf912, buf919, buf920, buf1327, buf929, buf936, buf937, buf944, buf945, buf1328, buf954, buf961, buf962, buf969, buf970, buf1329, buf979, buf986, buf987, buf994, buf995, buf1330, buf1003, buf1004, buf1011, buf1012, buf1331, buf1021, buf1028, buf1029, buf1036, buf1037, buf1332, buf1046, buf1053, buf1054, buf1061, buf1062, buf1333, buf1071, buf1078, buf1079, buf1086, buf1087, buf1334, buf1096, buf1103, buf1104, buf1111, buf1112, buf1335, buf1121, buf1128, buf1129, buf1136, buf1137, buf1336, buf1146, buf1153, buf1154, buf1161, buf1162, buf1337, buf1171, buf1178, buf1179, buf1186, buf1187, buf1338, buf1196, buf1203, buf1204, buf1211, buf1212, buf1339, buf1220, buf1221, buf1228, buf1229, buf1340, buf1238, buf1246, buf1255, buf1262, buf1268, buf1269, buf1276, buf1284, as_strided(buf1275, (1000, 1280), (1280, 1)), buf1341, buf1342, buf1343, buf1344, buf1345, buf1346, buf1347, buf1348, buf1349, buf1350, buf1351, buf1352, buf1353, buf1354, buf1355, buf1356, buf1357, buf1358, buf1359, buf1360, buf1361, buf1362, buf1363, buf1364, buf1365, buf1366, buf1367, buf1368, buf1369, buf1370, buf1371, buf1372, buf1373, buf1374, buf1375, s0, ) | |
NameError: name 's0' is not defined | |
TorchDynamo optimized model failed to run because of following error | |
cuda train mobilenet_v2_quantized_qat FAIL | |
Running torchbench.py mobilenet_v3_large... | |
ERROR:common:'int' object has no attribute 'size' | |
While executing %sym_size : [#users=10] = placeholder[target=sym_size] | |
Original traceback: | |
Module stack: {} | |
File "/data/users/ezyang/vision/torchvision/models/mobilenetv3.py", line 213, in _forward_impl | |
x = torch.flatten(x, 1) | |
| File "/data/users/ezyang/vision/torchvision/models/mobilenetv3.py", line 220, in forward | |
return self._forward_impl(x) | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1041, in check_accuracy | |
new_result = optimized_model_iter_fn(model, example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 945, in run_n_iterations | |
self.model_iter_fn(mod, inputs, collect_outputs=False) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 332, in forward_and_backward_pass | |
cloned_inputs = clone_inputs(inputs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 337, in <graph break in forward_and_backward_pass> | |
self.grad_scaler.scale(loss).backward() | |
File "/data/users/ezyang/pytorch-tmp/torch/_tensor.py", line 488, in backward | |
torch.autograd.backward( | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/__init__.py", line 197, in backward | |
Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/function.py", line 270, in apply | |
return user_fn(self, *args) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 499, in backward | |
CompiledFunction.compiled_bw = aot_config.bw_compiler( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/optimizations/backends.py", line 555, in _wrapped_bw_compiler | |
return disable(bw_compiler(*args, **kwargs)) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 362, in bw_compiler | |
return compile_fx_inner( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/debug_utils.py", line 466, in debug_wrapper | |
compiled_fn = compiler_fn(gm, example_inputs, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/debug.py", line 177, in inner | |
return fn(*args, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/contextlib.py", line 79, in inner | |
return func(*args, **kwds) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 122, in compile_fx_inner | |
graph.run(*example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 129, in run | |
return super().run(*args) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 130, in run | |
self.env[node] = self.run_node(node) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 299, in run_node | |
result = super().run_node(n) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 171, in run_node | |
return getattr(self, n.op)(n.target, args, kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 199, in placeholder | |
sizes, strides = self.static_sizes_strides(example) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 53, in static_sizes_strides | |
size = [sympy.Integer(i) for i in ex.size()] | |
AttributeError: 'int' object has no attribute 'size' | |
While executing %sym_size : [#users=10] = placeholder[target=sym_size] | |
Original traceback: | |
Module stack: {} | |
File "/data/users/ezyang/vision/torchvision/models/mobilenetv3.py", line 213, in _forward_impl | |
x = torch.flatten(x, 1) | |
| File "/data/users/ezyang/vision/torchvision/models/mobilenetv3.py", line 220, in forward | |
return self._forward_impl(x) | |
TorchDynamo optimized model failed to run because of following error | |
cuda train mobilenet_v3_large FAIL | |
devgpu001:632787:632787 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth0 | |
devgpu001:632787:632787 [0] NCCL INFO NCCL_SOCKET_IFNAME set to eth0 | |
devgpu001:632787:632787 [0] NCCL INFO Bootstrap : Using eth0:2803:6081:d0a8:baaf::1<0> | |
devgpu001:632787:632787 [0] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation | |
devgpu001:632787:632787 [0] NCCL INFO cudaDriverVersion 11040 | |
NCCL version 2.14.3+cuda11.4 | |
devgpu001:632787:633526 [0] NCCL INFO NCCL_IB_DISABLE set by environment to 1. | |
devgpu001:632787:633526 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth0 | |
devgpu001:632787:633526 [0] NCCL INFO NET/Socket : Using [0]eth0:2803:6081:d0a8:baaf::1<0> | |
devgpu001:632787:633526 [0] NCCL INFO Using network Socket | |
devgpu001:632787:633526 [0] NCCL INFO NET/Socket : GPU Direct RDMA Disabled for HCA 0 'eth0' | |
devgpu001:632787:633526 [0] NCCL INFO === System : maxBw 5000.0 totalBw 0.0 === | |
devgpu001:632787:633526 [0] NCCL INFO CPU/0 (1/1/2) | |
devgpu001:632787:633526 [0] NCCL INFO + PCI[12.0] - PCI/D000 (11f840001d9bfbe1) | |
devgpu001:632787:633526 [0] NCCL INFO + PCI[24.0] - PCI/F000 (11f840001d9bfbe0) | |
devgpu001:632787:633526 [0] NCCL INFO + PCI[24.0] - GPU/11000 (0) | |
devgpu001:632787:633526 [0] NCCL INFO + PCI[12.0] - NIC/30000 | |
devgpu001:632787:633526 [0] NCCL INFO ========================================== | |
devgpu001:632787:633526 [0] NCCL INFO GPU/11000 :GPU/11000 (0/5000.000000/LOC) CPU/0 (3/12.000000/PHB) | |
devgpu001:632787:633526 [0] NCCL INFO Setting affinity for GPU 0 to ffffff,00000000,00000000,00ffffff | |
devgpu001:632787:633526 [0] NCCL INFO Pattern 4, crossNic 0, nChannels 16, bw 44.000000/44.000000, type LOC/PIX, sameChannels 1 | |
devgpu001:632787:633526 [0] NCCL INFO 0 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 1 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 2 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 3 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 4 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 5 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 6 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 7 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 8 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 9 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 10 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 11 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 12 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 13 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 14 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 15 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO Pattern 3, crossNic 0, nChannels 16, bw 44.000000/44.000000, type LOC/PIX, sameChannels 1 | |
devgpu001:632787:633526 [0] NCCL INFO 0 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 1 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 2 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 3 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 4 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 5 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 6 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 7 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 8 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 9 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 10 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 11 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 12 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 13 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 14 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 15 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO Pattern 3, crossNic 0, nChannels 16, bw 44.000000/44.000000, type LOC/PIX, sameChannels 1 | |
devgpu001:632787:633526 [0] NCCL INFO 0 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 1 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 2 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 3 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 4 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 5 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 6 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 7 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 8 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 9 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 10 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 11 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 12 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 13 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 14 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO 15 : GPU/0 | |
devgpu001:632787:633526 [0] NCCL INFO Tree 0 : -1 -> 0 -> -1/-1/-1 | |
devgpu001:632787:633526 [0] NCCL INFO Tree 16 : -1 -> 0 -> -1/-1/-1 | |
devgpu001:632787:633526 [0] NCCL INFO Tree 1 : -1 -> 0 -> -1/-1/-1 | |
devgpu001:632787:633526 [0] NCCL INFO Tree 17 : -1 -> 0 -> -1/-1/-1 | |
devgpu001:632787:633526 [0] NCCL INFO Tree 2 : -1 -> 0 -> -1/-1/-1 | |
devgpu001:632787:633526 [0] NCCL INFO Tree 18 : -1 -> 0 -> -1/-1/-1 | |
devgpu001:632787:633526 [0] NCCL INFO Tree 3 : -1 -> 0 -> -1/-1/-1 | |
devgpu001:632787:633526 [0] NCCL INFO Tree 19 : -1 -> 0 -> -1/-1/-1 | |
devgpu001:632787:633526 [0] NCCL INFO Tree 4 : -1 -> 0 -> -1/-1/-1 | |
devgpu001:632787:633526 [0] NCCL INFO Tree 20 : -1 -> 0 -> -1/-1/-1 | |
devgpu001:632787:633526 [0] NCCL INFO Tree 5 : -1 -> 0 -> -1/-1/-1 | |
devgpu001:632787:633526 [0] NCCL INFO Tree 21 : -1 -> 0 -> -1/-1/-1 | |
devgpu001:632787:633526 [0] NCCL INFO Tree 6 : -1 -> 0 -> -1/-1/-1 | |
devgpu001:632787:633526 [0] NCCL INFO Tree 22 : -1 -> 0 -> -1/-1/-1 | |
devgpu001:632787:633526 [0] NCCL INFO Tree 7 : -1 -> 0 -> -1/-1/-1 | |
devgpu001:632787:633526 [0] NCCL INFO Tree 23 : -1 -> 0 -> -1/-1/-1 | |
devgpu001:632787:633526 [0] NCCL INFO Tree 8 : -1 -> 0 -> -1/-1/-1 | |
devgpu001:632787:633526 [0] NCCL INFO Tree 24 : -1 -> 0 -> -1/-1/-1 | |
devgpu001:632787:633526 [0] NCCL INFO Tree 9 : -1 -> 0 -> -1/-1/-1 | |
devgpu001:632787:633526 [0] NCCL INFO Tree 25 : -1 -> 0 -> -1/-1/-1 | |
devgpu001:632787:633526 [0] NCCL INFO Tree 10 : -1 -> 0 -> -1/-1/-1 | |
devgpu001:632787:633526 [0] NCCL INFO Tree 26 : -1 -> 0 -> -1/-1/-1 | |
devgpu001:632787:633526 [0] NCCL INFO Tree 11 : -1 -> 0 -> -1/-1/-1 | |
devgpu001:632787:633526 [0] NCCL INFO Tree 27 : -1 -> 0 -> -1/-1/-1 | |
devgpu001:632787:633526 [0] NCCL INFO Tree 12 : -1 -> 0 -> -1/-1/-1 | |
devgpu001:632787:633526 [0] NCCL INFO Tree 28 : -1 -> 0 -> -1/-1/-1 | |
devgpu001:632787:633526 [0] NCCL INFO Tree 13 : -1 -> 0 -> -1/-1/-1 | |
devgpu001:632787:633526 [0] NCCL INFO Tree 29 : -1 -> 0 -> -1/-1/-1 | |
devgpu001:632787:633526 [0] NCCL INFO Tree 14 : -1 -> 0 -> -1/-1/-1 | |
devgpu001:632787:633526 [0] NCCL INFO Tree 30 : -1 -> 0 -> -1/-1/-1 | |
devgpu001:632787:633526 [0] NCCL INFO Tree 15 : -1 -> 0 -> -1/-1/-1 | |
devgpu001:632787:633526 [0] NCCL INFO Tree 31 : -1 -> 0 -> -1/-1/-1 | |
devgpu001:632787:633526 [0] NCCL INFO Channel 00/32 : 0 | |
devgpu001:632787:633526 [0] NCCL INFO Channel 01/32 : 0 | |
devgpu001:632787:633526 [0] NCCL INFO Channel 02/32 : 0 | |
devgpu001:632787:633526 [0] NCCL INFO Channel 03/32 : 0 | |
devgpu001:632787:633526 [0] NCCL INFO Channel 04/32 : 0 | |
devgpu001:632787:633526 [0] NCCL INFO Channel 05/32 : 0 | |
devgpu001:632787:633526 [0] NCCL INFO Channel 06/32 : 0 | |
devgpu001:632787:633526 [0] NCCL INFO Channel 07/32 : 0 | |
devgpu001:632787:633526 [0] NCCL INFO Channel 08/32 : 0 | |
devgpu001:632787:633526 [0] NCCL INFO Channel 09/32 : 0 | |
devgpu001:632787:633526 [0] NCCL INFO Channel 10/32 : 0 | |
devgpu001:632787:633526 [0] NCCL INFO Channel 11/32 : 0 | |
devgpu001:632787:633526 [0] NCCL INFO Channel 12/32 : 0 | |
devgpu001:632787:633526 [0] NCCL INFO Channel 13/32 : 0 | |
devgpu001:632787:633526 [0] NCCL INFO Channel 14/32 : 0 | |
devgpu001:632787:633526 [0] NCCL INFO Channel 15/32 : 0 | |
devgpu001:632787:633526 [0] NCCL INFO Channel 16/32 : 0 | |
devgpu001:632787:633526 [0] NCCL INFO Channel 17/32 : 0 | |
devgpu001:632787:633526 [0] NCCL INFO Channel 18/32 : 0 | |
devgpu001:632787:633526 [0] NCCL INFO Channel 19/32 : 0 | |
devgpu001:632787:633526 [0] NCCL INFO Channel 20/32 : 0 | |
devgpu001:632787:633526 [0] NCCL INFO Channel 21/32 : 0 | |
devgpu001:632787:633526 [0] NCCL INFO Channel 22/32 : 0 | |
devgpu001:632787:633526 [0] NCCL INFO Channel 23/32 : 0 | |
devgpu001:632787:633526 [0] NCCL INFO Channel 24/32 : 0 | |
devgpu001:632787:633526 [0] NCCL INFO Channel 25/32 : 0 | |
devgpu001:632787:633526 [0] NCCL INFO Channel 26/32 : 0 | |
devgpu001:632787:633526 [0] NCCL INFO Channel 27/32 : 0 | |
devgpu001:632787:633526 [0] NCCL INFO Channel 28/32 : 0 | |
devgpu001:632787:633526 [0] NCCL INFO Channel 29/32 : 0 | |
devgpu001:632787:633526 [0] NCCL INFO Channel 30/32 : 0 | |
devgpu001:632787:633526 [0] NCCL INFO Channel 31/32 : 0 | |
devgpu001:632787:633526 [0] NCCL INFO Ring 00 : 0 -> 0 -> 0 | |
devgpu001:632787:633526 [0] NCCL INFO Ring 01 : 0 -> 0 -> 0 | |
devgpu001:632787:633526 [0] NCCL INFO Ring 02 : 0 -> 0 -> 0 | |
devgpu001:632787:633526 [0] NCCL INFO Ring 03 : 0 -> 0 -> 0 | |
devgpu001:632787:633526 [0] NCCL INFO Ring 04 : 0 -> 0 -> 0 | |
devgpu001:632787:633526 [0] NCCL INFO Ring 05 : 0 -> 0 -> 0 | |
devgpu001:632787:633526 [0] NCCL INFO Ring 06 : 0 -> 0 -> 0 | |
devgpu001:632787:633526 [0] NCCL INFO Ring 07 : 0 -> 0 -> 0 | |
devgpu001:632787:633526 [0] NCCL INFO Ring 08 : 0 -> 0 -> 0 | |
devgpu001:632787:633526 [0] NCCL INFO Ring 09 : 0 -> 0 -> 0 | |
devgpu001:632787:633526 [0] NCCL INFO Ring 10 : 0 -> 0 -> 0 | |
devgpu001:632787:633526 [0] NCCL INFO Ring 11 : 0 -> 0 -> 0 | |
devgpu001:632787:633526 [0] NCCL INFO Ring 12 : 0 -> 0 -> 0 | |
devgpu001:632787:633526 [0] NCCL INFO Ring 13 : 0 -> 0 -> 0 | |
devgpu001:632787:633526 [0] NCCL INFO Ring 14 : 0 -> 0 -> 0 | |
devgpu001:632787:633526 [0] NCCL INFO Ring 15 : 0 -> 0 -> 0 | |
devgpu001:632787:633526 [0] NCCL INFO Ring 16 : 0 -> 0 -> 0 | |
devgpu001:632787:633526 [0] NCCL INFO Ring 17 : 0 -> 0 -> 0 | |
devgpu001:632787:633526 [0] NCCL INFO Ring 18 : 0 -> 0 -> 0 | |
devgpu001:632787:633526 [0] NCCL INFO Ring 19 : 0 -> 0 -> 0 | |
devgpu001:632787:633526 [0] NCCL INFO Ring 20 : 0 -> 0 -> 0 | |
devgpu001:632787:633526 [0] NCCL INFO Ring 21 : 0 -> 0 -> 0 | |
devgpu001:632787:633526 [0] NCCL INFO Ring 22 : 0 -> 0 -> 0 | |
devgpu001:632787:633526 [0] NCCL INFO Ring 23 : 0 -> 0 -> 0 | |
devgpu001:632787:633526 [0] NCCL INFO Ring 24 : 0 -> 0 -> 0 | |
devgpu001:632787:633526 [0] NCCL INFO Ring 25 : 0 -> 0 -> 0 | |
devgpu001:632787:633526 [0] NCCL INFO Ring 26 : 0 -> 0 -> 0 | |
devgpu001:632787:633526 [0] NCCL INFO Ring 27 : 0 -> 0 -> 0 | |
devgpu001:632787:633526 [0] NCCL INFO Ring 28 : 0 -> 0 -> 0 | |
devgpu001:632787:633526 [0] NCCL INFO Ring 29 : 0 -> 0 -> 0 | |
devgpu001:632787:633526 [0] NCCL INFO Ring 30 : 0 -> 0 -> 0 | |
devgpu001:632787:633526 [0] NCCL INFO Ring 31 : 0 -> 0 -> 0 | |
devgpu001:632787:633526 [0] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 | |
devgpu001:632787:633526 [0] NCCL INFO Connected all rings | |
devgpu001:632787:633526 [0] NCCL INFO Connected all trees | |
devgpu001:632787:633526 [0] NCCL INFO 32 coll channels, 32 p2p channels, 32 p2p channels per peer | |
devgpu001:632787:633534 [0] NCCL INFO New proxy send connection 0 from local rank 0, transport 2 | |
devgpu001:632787:633526 [0] NCCL INFO Connection to proxy localRank 0 -> connection 0x7f9820002e80 | |
devgpu001:632787:633526 [0] NCCL INFO comm 0x564b425bfa90 rank 0 nranks 1 cudaDev 0 busId 11000 - Init COMPLETE | |
Running torchbench.py moco... | |
ERROR:common:argument of type: <class 'range_iterator'> | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1041, in check_accuracy | |
new_result = optimized_model_iter_fn(model, example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 945, in run_n_iterations | |
self.model_iter_fn(mod, inputs, collect_outputs=False) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 332, in forward_and_backward_pass | |
cloned_inputs = clone_inputs(inputs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 335, in <graph break in forward_and_backward_pass> | |
pred = mod(*cloned_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/parallel/distributed.py", line 1093, in forward | |
output = self._run_ddp_forward(*inputs, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/parallel/distributed.py", line 1047, in _run_ddp_forward | |
return module_to_run(*inputs[0], **kwargs[0]) | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "/data/users/ezyang/benchmark/torchbenchmark/models/moco/moco/builder.py", line 130, in forward | |
self._momentum_update_key_encoder() # update the key encoder | |
File "/data/users/ezyang/benchmark/torchbenchmark/models/moco/moco/builder.py", line 133, in <graph break in forward> | |
im_k, idx_unshuffle = self._batch_shuffle_ddp(im_k) | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/grad_mode.py", line 27, in decorate_context | |
return func(*args, **kwargs) | |
File "/data/users/ezyang/benchmark/torchbenchmark/models/moco/moco/builder.py", line 76, in _batch_shuffle_ddp | |
x_gather = concat_all_gather(x) | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/grad_mode.py", line 27, in decorate_context | |
return func(*args, **kwargs) | |
File "/data/users/ezyang/benchmark/torchbenchmark/models/moco/moco/builder.py", line 164, in concat_all_gather | |
@torch.no_grad() | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 893, in forward | |
return compiled_f( | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 880, in new_func | |
compiled_fn = create_aot_dispatcher_function( | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 603, in create_aot_dispatcher_function | |
return aot_dispatch_base(flat_fn, fake_flat_tensor_args, aot_config) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 326, in aot_dispatch_base | |
fw_module = make_fx(flat_fn, aot_config.decompositions)(*flat_args) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/experimental/proxy_tensor.py", line 671, in wrapped | |
t = dispatch_trace(wrap_key(func, args, fx_tracer), tracer=fx_tracer, concrete_args=tuple(phs)) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/experimental/proxy_tensor.py", line 422, in dispatch_trace | |
graph = tracer.trace(root, concrete_args) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/_symbolic_trace.py", line 739, in trace | |
(self.create_arg(fn(*args)),), | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/experimental/proxy_tensor.py", line 412, in create_arg | |
return super().create_arg(a) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/_symbolic_trace.py", line 344, in create_arg | |
return super().create_arg(a) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/proxy.py", line 140, in create_arg | |
return type(a)(self.create_arg(elem) for elem in a) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/proxy.py", line 140, in <genexpr> | |
return type(a)(self.create_arg(elem) for elem in a) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/experimental/proxy_tensor.py", line 412, in create_arg | |
return super().create_arg(a) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/_symbolic_trace.py", line 344, in create_arg | |
return super().create_arg(a) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/proxy.py", line 165, in create_arg | |
raise NotImplementedError(f"argument of type: {type(a)}") | |
NotImplementedError: argument of type: <class 'range_iterator'> | |
incomplete graph: | |
class <lambda>(torch.nn.Module): | |
def forward(self): | |
pass | |
TorchDynamo optimized model failed to run because of following error | |
cuda train moco FAIL | |
Running torchbench.py nvidia_deeprecommender... | |
cuda train nvidia_deeprecommender PASS | |
Running torchbench.py pytorch_CycleGAN_and_pix2pix... | |
ERROR:common:'int' object has no attribute 'size' | |
While executing %sym_size : [#users=102] = placeholder[target=sym_size] | |
Original traceback: | |
Module stack: {'self_model': <class 'torch.nn.modules.container.Sequential'>, 'self_model_2': <class 'torch.nn.modules.instancenorm.InstanceNorm2d'>} | |
File "/data/users/ezyang/benchmark/torchbenchmark/models/pytorch_CycleGAN_and_pix2pix/models/networks.py", line 372, in forward | |
return self.model(input) | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1041, in check_accuracy | |
new_result = optimized_model_iter_fn(model, example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 945, in run_n_iterations | |
self.model_iter_fn(mod, inputs, collect_outputs=False) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 332, in forward_and_backward_pass | |
cloned_inputs = clone_inputs(inputs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 337, in <graph break in forward_and_backward_pass> | |
self.grad_scaler.scale(loss).backward() | |
File "/data/users/ezyang/pytorch-tmp/torch/_tensor.py", line 488, in backward | |
torch.autograd.backward( | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/__init__.py", line 197, in backward | |
Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/function.py", line 270, in apply | |
return user_fn(self, *args) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 499, in backward | |
CompiledFunction.compiled_bw = aot_config.bw_compiler( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/optimizations/backends.py", line 555, in _wrapped_bw_compiler | |
return disable(bw_compiler(*args, **kwargs)) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 362, in bw_compiler | |
return compile_fx_inner( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/debug_utils.py", line 466, in debug_wrapper | |
compiled_fn = compiler_fn(gm, example_inputs, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/debug.py", line 177, in inner | |
return fn(*args, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/contextlib.py", line 79, in inner | |
return func(*args, **kwds) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 122, in compile_fx_inner | |
graph.run(*example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 129, in run | |
return super().run(*args) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 130, in run | |
self.env[node] = self.run_node(node) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 299, in run_node | |
result = super().run_node(n) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 171, in run_node | |
return getattr(self, n.op)(n.target, args, kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 199, in placeholder | |
sizes, strides = self.static_sizes_strides(example) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 53, in static_sizes_strides | |
size = [sympy.Integer(i) for i in ex.size()] | |
AttributeError: 'int' object has no attribute 'size' | |
While executing %sym_size : [#users=102] = placeholder[target=sym_size] | |
Original traceback: | |
Module stack: {'self_model': <class 'torch.nn.modules.container.Sequential'>, 'self_model_2': <class 'torch.nn.modules.instancenorm.InstanceNorm2d'>} | |
File "/data/users/ezyang/benchmark/torchbenchmark/models/pytorch_CycleGAN_and_pix2pix/models/networks.py", line 372, in forward | |
return self.model(input) | |
--dataroot /data/users/ezyang/benchmark/torchbenchmark/data/.data/pytorch_CycleGAN_and_pix2pix_inputs/datasets/horse2zebra --name horse2zebra --model cycle_gan --display_id 0 --n_epochs 3 --n_epochs_decay 3 --gpu_ids 0 --checkpoints_dir /data/users/ezyang/benchmark/torchbenchmark/models/pytorch_CycleGAN_and_pix2pix/.data/checkpoints | |
TorchDynamo optimized model failed to run because of following error | |
cuda train pytorch_CycleGAN_and_pix2pix FAIL | |
Running torchbench.py pytorch_stargan... | |
ERROR:common:'int' object has no attribute 'size' | |
While executing %sym_size_4 : [#users=4] = placeholder[target=sym_size_4] | |
Original traceback: | |
Module stack: {'self_main': <class 'torch.nn.modules.container.Sequential'>, 'self_main_1': <class 'torch.nn.modules.instancenorm.InstanceNorm2d'>} | |
File "/data/users/ezyang/benchmark/torchbenchmark/models/pytorch_stargan/model.py", line 64, in forward | |
return self.main(x) | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1041, in check_accuracy | |
new_result = optimized_model_iter_fn(model, example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 945, in run_n_iterations | |
self.model_iter_fn(mod, inputs, collect_outputs=False) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 332, in forward_and_backward_pass | |
cloned_inputs = clone_inputs(inputs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 337, in <graph break in forward_and_backward_pass> | |
self.grad_scaler.scale(loss).backward() | |
File "/data/users/ezyang/pytorch-tmp/torch/_tensor.py", line 488, in backward | |
torch.autograd.backward( | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/__init__.py", line 197, in backward | |
Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/function.py", line 270, in apply | |
return user_fn(self, *args) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 499, in backward | |
CompiledFunction.compiled_bw = aot_config.bw_compiler( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/optimizations/backends.py", line 555, in _wrapped_bw_compiler | |
return disable(bw_compiler(*args, **kwargs)) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 362, in bw_compiler | |
return compile_fx_inner( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/debug_utils.py", line 466, in debug_wrapper | |
compiled_fn = compiler_fn(gm, example_inputs, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/debug.py", line 177, in inner | |
return fn(*args, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/contextlib.py", line 79, in inner | |
return func(*args, **kwds) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 122, in compile_fx_inner | |
graph.run(*example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 129, in run | |
return super().run(*args) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 130, in run | |
self.env[node] = self.run_node(node) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 299, in run_node | |
result = super().run_node(n) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 171, in run_node | |
return getattr(self, n.op)(n.target, args, kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 199, in placeholder | |
sizes, strides = self.static_sizes_strides(example) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 53, in static_sizes_strides | |
size = [sympy.Integer(i) for i in ex.size()] | |
AttributeError: 'int' object has no attribute 'size' | |
While executing %sym_size_4 : [#users=4] = placeholder[target=sym_size_4] | |
Original traceback: | |
Module stack: {'self_main': <class 'torch.nn.modules.container.Sequential'>, 'self_main_1': <class 'torch.nn.modules.instancenorm.InstanceNorm2d'>} | |
File "/data/users/ezyang/benchmark/torchbenchmark/models/pytorch_stargan/model.py", line 64, in forward | |
return self.main(x) | |
TorchDynamo optimized model failed to run because of following error | |
cuda train pytorch_stargan FAIL | |
Running torchbench.py pytorch_struct... | |
ERROR:common:'int' object has no attribute 'size' | |
While executing %sym_size : [#users=2] = placeholder[target=sym_size] | |
Original traceback: | |
Module stack: {} | |
File "/data/users/ezyang/benchmark/torchbenchmark/models/pytorch_struct/networks/NeuralCFG.py", line 49, in terms | |
torch.einsum("vh,th->tv", self.word_emb, self.mlp1(self.term_emb)) | |
| File "/data/users/ezyang/benchmark/torchbenchmark/models/pytorch_struct/networks/NeuralCFG.py", line 77, in forward | |
return terms(input), rules(batch), roots(batch) | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1041, in check_accuracy | |
new_result = optimized_model_iter_fn(model, example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 945, in run_n_iterations | |
self.model_iter_fn(mod, inputs, collect_outputs=False) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 332, in forward_and_backward_pass | |
cloned_inputs = clone_inputs(inputs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 337, in <graph break in forward_and_backward_pass> | |
self.grad_scaler.scale(loss).backward() | |
File "/data/users/ezyang/pytorch-tmp/torch/_tensor.py", line 488, in backward | |
torch.autograd.backward( | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/__init__.py", line 197, in backward | |
Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/function.py", line 270, in apply | |
return user_fn(self, *args) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 499, in backward | |
CompiledFunction.compiled_bw = aot_config.bw_compiler( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/optimizations/backends.py", line 555, in _wrapped_bw_compiler | |
return disable(bw_compiler(*args, **kwargs)) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 362, in bw_compiler | |
return compile_fx_inner( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/debug_utils.py", line 466, in debug_wrapper | |
compiled_fn = compiler_fn(gm, example_inputs, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/debug.py", line 177, in inner | |
return fn(*args, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/contextlib.py", line 79, in inner | |
return func(*args, **kwds) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 122, in compile_fx_inner | |
graph.run(*example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 129, in run | |
return super().run(*args) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 130, in run | |
self.env[node] = self.run_node(node) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 299, in run_node | |
result = super().run_node(n) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 171, in run_node | |
return getattr(self, n.op)(n.target, args, kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 199, in placeholder | |
sizes, strides = self.static_sizes_strides(example) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 53, in static_sizes_strides | |
size = [sympy.Integer(i) for i in ex.size()] | |
AttributeError: 'int' object has no attribute 'size' | |
While executing %sym_size : [#users=2] = placeholder[target=sym_size] | |
Original traceback: | |
Module stack: {} | |
File "/data/users/ezyang/benchmark/torchbenchmark/models/pytorch_struct/networks/NeuralCFG.py", line 49, in terms | |
torch.einsum("vh,th->tv", self.word_emb, self.mlp1(self.term_emb)) | |
| File "/data/users/ezyang/benchmark/torchbenchmark/models/pytorch_struct/networks/NeuralCFG.py", line 77, in forward | |
return terms(input), rules(batch), roots(batch) | |
TorchDynamo optimized model failed to run because of following error | |
cuda train pytorch_struct FAIL | |
Running torchbench.py pytorch_unet... | |
ERROR:common:RuntimeError: Overloaded torch operator invoked from Python failed to many any schema: | |
aten::arange() expected at most 5 argument(s) but received 7 argument(s). Declaration: aten::arange(Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() expected at most 6 argument(s) but received 7 argument(s). Declaration: aten::arange.start(Scalar start, Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() Expected a value of type 'number' for argument 'end' but instead found type 'Float'. | |
Position: 1 | |
Value: 80.0000000000000 | |
Declaration: aten::arange.start_step(Scalar start, Scalar end, Scalar step=1, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
Cast error details: Cannot cast 80.0000000000000 to number | |
aten::arange() expected at most 4 argument(s) but received 7 argument(s). Declaration: aten::arange.start_out(Scalar start, Scalar end, Scalar step=1, *, Tensor(a!) out) -> Tensor(a!) | |
aten::arange() expected at most 2 argument(s) but received 7 argument(s). Declaration: aten::arange.out(Scalar end, *, Tensor(a!) out) -> Tensor(a!) | |
target: aten.arange.default | |
args[0]: 80.0000000000000 | |
kwargs: {'dtype': torch.float32, 'device': device(type='cuda', index=0), 'pin_memory': False} | |
While executing %arange : [#users=1] = call_function[target=torch.ops.aten.arange.default](args = (%mul_30,), kwargs = {dtype: torch.float32, device: cuda:0, pin_memory: False}) | |
Original traceback: | |
Module stack: {'self_up1': <class 'torchbenchmark.models.pytorch_unet.pytorch_unet.unet.unet_parts.Up'>, 'self_up1_up': <class 'torch.nn.modules.upsampling.Upsample'>} | |
File "/data/users/ezyang/benchmark/torchbenchmark/models/pytorch_unet/pytorch_unet/unet/unet_parts.py", line 57, in forward | |
x1 = self.up(x1) | |
| File "/data/users/ezyang/benchmark/torchbenchmark/models/pytorch_unet/pytorch_unet/unet/unet_model.py", line 31, in forward | |
x = self.up1(x5, x4) | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 239, in call_function | |
out = lowerings[target](*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/lowering.py", line 204, in wrapped | |
return decomp_fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/lowering.py", line 1194, in arange | |
return fallback_arange( | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/lowering.py", line 893, in handler | |
result = ir.FallbackKernel.create(kernel, *args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/ir.py", line 2890, in create | |
example_output = kernel( | |
File "/data/users/ezyang/pytorch-tmp/torch/_ops.py", line 446, in __call__ | |
return self._op(*args, **kwargs or {}) | |
RuntimeError: Overloaded torch operator invoked from Python failed to many any schema: | |
aten::arange() expected at most 5 argument(s) but received 7 argument(s). Declaration: aten::arange(Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() expected at most 6 argument(s) but received 7 argument(s). Declaration: aten::arange.start(Scalar start, Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() Expected a value of type 'number' for argument 'end' but instead found type 'Float'. | |
Position: 1 | |
Value: 80.0000000000000 | |
Declaration: aten::arange.start_step(Scalar start, Scalar end, Scalar step=1, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
Cast error details: Cannot cast 80.0000000000000 to number | |
aten::arange() expected at most 4 argument(s) but received 7 argument(s). Declaration: aten::arange.start_out(Scalar start, Scalar end, Scalar step=1, *, Tensor(a!) out) -> Tensor(a!) | |
aten::arange() expected at most 2 argument(s) but received 7 argument(s). Declaration: aten::arange.out(Scalar end, *, Tensor(a!) out) -> Tensor(a!) | |
The above exception was the direct cause of the following exception: | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1041, in check_accuracy | |
new_result = optimized_model_iter_fn(model, example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 945, in run_n_iterations | |
self.model_iter_fn(mod, inputs, collect_outputs=False) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 332, in forward_and_backward_pass | |
cloned_inputs = clone_inputs(inputs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 335, in <graph break in forward_and_backward_pass> | |
pred = mod(*cloned_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "/data/users/ezyang/benchmark/torchbenchmark/models/pytorch_unet/pytorch_unet/unet/unet_model.py", line 25, in forward | |
def forward(self, x): | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 893, in forward | |
return compiled_f( | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 880, in new_func | |
compiled_fn = create_aot_dispatcher_function( | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 600, in create_aot_dispatcher_function | |
aot_dispatch_autograd(flat_fn, fake_flat_tensor_args, aot_config) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 459, in aot_dispatch_autograd | |
compiled_fw_func = aot_config.fw_compiler(fw_module, deduped_flat_args) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 351, in fw_compiler | |
return compile_fx_inner( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/debug_utils.py", line 466, in debug_wrapper | |
compiled_fn = compiler_fn(gm, example_inputs, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/debug.py", line 177, in inner | |
return fn(*args, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/contextlib.py", line 79, in inner | |
return func(*args, **kwds) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 122, in compile_fx_inner | |
graph.run(*example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 129, in run | |
return super().run(*args) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 130, in run | |
self.env[node] = self.run_node(node) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 299, in run_node | |
result = super().run_node(n) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 171, in run_node | |
return getattr(self, n.op)(n.target, args, kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 242, in call_function | |
raise LoweringException(e, target, args, kwargs) from e | |
torch._inductor.exc.LoweringException: RuntimeError: Overloaded torch operator invoked from Python failed to many any schema: | |
aten::arange() expected at most 5 argument(s) but received 7 argument(s). Declaration: aten::arange(Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() expected at most 6 argument(s) but received 7 argument(s). Declaration: aten::arange.start(Scalar start, Scalar end, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
aten::arange() Expected a value of type 'number' for argument 'end' but instead found type 'Float'. | |
Position: 1 | |
Value: 80.0000000000000 | |
Declaration: aten::arange.start_step(Scalar start, Scalar end, Scalar step=1, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor | |
Cast error details: Cannot cast 80.0000000000000 to number | |
aten::arange() expected at most 4 argument(s) but received 7 argument(s). Declaration: aten::arange.start_out(Scalar start, Scalar end, Scalar step=1, *, Tensor(a!) out) -> Tensor(a!) | |
aten::arange() expected at most 2 argument(s) but received 7 argument(s). Declaration: aten::arange.out(Scalar end, *, Tensor(a!) out) -> Tensor(a!) | |
target: aten.arange.default | |
args[0]: 80.0000000000000 | |
kwargs: {'dtype': torch.float32, 'device': device(type='cuda', index=0), 'pin_memory': False} | |
While executing %arange : [#users=1] = call_function[target=torch.ops.aten.arange.default](args = (%mul_30,), kwargs = {dtype: torch.float32, device: cuda:0, pin_memory: False}) | |
Original traceback: | |
Module stack: {'self_up1': <class 'torchbenchmark.models.pytorch_unet.pytorch_unet.unet.unet_parts.Up'>, 'self_up1_up': <class 'torch.nn.modules.upsampling.Upsample'>} | |
File "/data/users/ezyang/benchmark/torchbenchmark/models/pytorch_unet/pytorch_unet/unet/unet_parts.py", line 57, in forward | |
x1 = self.up(x1) | |
| File "/data/users/ezyang/benchmark/torchbenchmark/models/pytorch_unet/pytorch_unet/unet/unet_model.py", line 31, in forward | |
x = self.up1(x5, x4) | |
TorchDynamo optimized model failed to run because of following error | |
cuda train pytorch_unet FAIL | |
Running torchbench.py resnet18... | |
ERROR:common:'int' object has no attribute 'size' | |
While executing %sym_size : [#users=2] = placeholder[target=sym_size] | |
Original traceback: | |
Module stack: {} | |
File "/data/users/ezyang/vision/torchvision/models/resnet.py", line 279, in _forward_impl | |
x = torch.flatten(x, 1) | |
| File "/data/users/ezyang/vision/torchvision/models/resnet.py", line 285, in forward | |
return self._forward_impl(x) | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1041, in check_accuracy | |
new_result = optimized_model_iter_fn(model, example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 945, in run_n_iterations | |
self.model_iter_fn(mod, inputs, collect_outputs=False) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 332, in forward_and_backward_pass | |
cloned_inputs = clone_inputs(inputs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 337, in <graph break in forward_and_backward_pass> | |
self.grad_scaler.scale(loss).backward() | |
File "/data/users/ezyang/pytorch-tmp/torch/_tensor.py", line 488, in backward | |
torch.autograd.backward( | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/__init__.py", line 197, in backward | |
Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/function.py", line 270, in apply | |
return user_fn(self, *args) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 499, in backward | |
CompiledFunction.compiled_bw = aot_config.bw_compiler( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/optimizations/backends.py", line 555, in _wrapped_bw_compiler | |
return disable(bw_compiler(*args, **kwargs)) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 362, in bw_compiler | |
return compile_fx_inner( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/debug_utils.py", line 466, in debug_wrapper | |
compiled_fn = compiler_fn(gm, example_inputs, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/debug.py", line 177, in inner | |
return fn(*args, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/contextlib.py", line 79, in inner | |
return func(*args, **kwds) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 122, in compile_fx_inner | |
graph.run(*example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 129, in run | |
return super().run(*args) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 130, in run | |
self.env[node] = self.run_node(node) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 299, in run_node | |
result = super().run_node(n) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 171, in run_node | |
return getattr(self, n.op)(n.target, args, kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 199, in placeholder | |
sizes, strides = self.static_sizes_strides(example) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 53, in static_sizes_strides | |
size = [sympy.Integer(i) for i in ex.size()] | |
AttributeError: 'int' object has no attribute 'size' | |
While executing %sym_size : [#users=2] = placeholder[target=sym_size] | |
Original traceback: | |
Module stack: {} | |
File "/data/users/ezyang/vision/torchvision/models/resnet.py", line 279, in _forward_impl | |
x = torch.flatten(x, 1) | |
| File "/data/users/ezyang/vision/torchvision/models/resnet.py", line 285, in forward | |
return self._forward_impl(x) | |
TorchDynamo optimized model failed to run because of following error | |
cuda train resnet18 FAIL | |
Running torchbench.py resnet50... | |
ERROR:common:'int' object has no attribute 'size' | |
While executing %sym_size : [#users=2] = placeholder[target=sym_size] | |
Original traceback: | |
Module stack: {} | |
File "/data/users/ezyang/vision/torchvision/models/resnet.py", line 279, in _forward_impl | |
x = torch.flatten(x, 1) | |
| File "/data/users/ezyang/vision/torchvision/models/resnet.py", line 285, in forward | |
return self._forward_impl(x) | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1041, in check_accuracy | |
new_result = optimized_model_iter_fn(model, example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 945, in run_n_iterations | |
self.model_iter_fn(mod, inputs, collect_outputs=False) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 332, in forward_and_backward_pass | |
cloned_inputs = clone_inputs(inputs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 337, in <graph break in forward_and_backward_pass> | |
self.grad_scaler.scale(loss).backward() | |
File "/data/users/ezyang/pytorch-tmp/torch/_tensor.py", line 488, in backward | |
torch.autograd.backward( | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/__init__.py", line 197, in backward | |
Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/function.py", line 270, in apply | |
return user_fn(self, *args) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 499, in backward | |
CompiledFunction.compiled_bw = aot_config.bw_compiler( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/optimizations/backends.py", line 555, in _wrapped_bw_compiler | |
return disable(bw_compiler(*args, **kwargs)) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 362, in bw_compiler | |
return compile_fx_inner( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/debug_utils.py", line 466, in debug_wrapper | |
compiled_fn = compiler_fn(gm, example_inputs, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/debug.py", line 177, in inner | |
return fn(*args, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/contextlib.py", line 79, in inner | |
return func(*args, **kwds) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 122, in compile_fx_inner | |
graph.run(*example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 129, in run | |
return super().run(*args) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 130, in run | |
self.env[node] = self.run_node(node) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 299, in run_node | |
result = super().run_node(n) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 171, in run_node | |
return getattr(self, n.op)(n.target, args, kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 199, in placeholder | |
sizes, strides = self.static_sizes_strides(example) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 53, in static_sizes_strides | |
size = [sympy.Integer(i) for i in ex.size()] | |
AttributeError: 'int' object has no attribute 'size' | |
While executing %sym_size : [#users=2] = placeholder[target=sym_size] | |
Original traceback: | |
Module stack: {} | |
File "/data/users/ezyang/vision/torchvision/models/resnet.py", line 279, in _forward_impl | |
x = torch.flatten(x, 1) | |
| File "/data/users/ezyang/vision/torchvision/models/resnet.py", line 285, in forward | |
return self._forward_impl(x) | |
TorchDynamo optimized model failed to run because of following error | |
cuda train resnet50 FAIL | |
Running torchbench.py resnet50_quantized_qat... | |
WARNING:common:fp64 golden ref were not generated for resnet50_quantized_qat | |
[2022-10-30 21:35:27,833] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:27,839] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:27,854] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:27,869] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:27,873] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:27,886] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:27,891] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:27,903] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:27,907] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:27,919] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:27,923] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:27,935] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:27,937] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:27,941] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:27,953] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:27,956] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:27,968] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:27,971] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:27,983] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:27,985] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:27,989] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,001] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,005] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,017] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,021] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,033] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,035] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,039] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,053] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,057] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,071] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,075] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,088] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,091] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,103] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,105] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,109] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,121] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,125] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,138] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,141] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,153] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,155] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,159] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,171] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,174] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,186] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,189] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,201] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,203] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,207] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,219] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,222] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,235] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,238] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,250] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,252] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,256] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,268] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,272] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,286] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,289] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,301] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,305] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,316] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,318] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,322] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,335] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,339] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,351] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,354] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,366] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,368] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,372] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,383] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,387] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,399] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,402] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,414] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,416] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,420] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,432] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,436] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,448] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,451] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,462] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,464] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,467] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,479] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,483] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,494] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,498] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,509] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,511] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,515] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,529] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,532] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,545] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,549] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,561] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,564] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,568] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,581] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,586] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,600] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,604] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,617] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,620] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,633] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,635] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,640] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,653] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,657] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,669] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,672] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,683] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,685] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,689] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,701] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,705] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,717] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,720] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,732] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,735] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,738] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,740] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,741] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
[2022-10-30 21:35:28,742] torch._inductor.ir: [WARNING] Using FallbackKernel: aten._fused_moving_avg_obs_fq_helper_functional | |
ERROR:common:name 's0' is not defined | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1041, in check_accuracy | |
new_result = optimized_model_iter_fn(model, example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 945, in run_n_iterations | |
self.model_iter_fn(mod, inputs, collect_outputs=False) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 332, in forward_and_backward_pass | |
cloned_inputs = clone_inputs(inputs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 335, in <graph break in forward_and_backward_pass> | |
pred = mod(*cloned_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/graph_module.py", line 660, in call_wrapped | |
return self._wrapped_call(self, *args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/graph_module.py", line 279, in __call__ | |
raise e | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/graph_module.py", line 269, in __call__ | |
return super(self.cls, obj).__call__(*args, **kwargs) # type: ignore[misc] | |
File "/data/users/ezyang/pytorch-tmp/torch/nn/modules/module.py", line 1423, in _call_impl | |
return forward_call(*input, **kwargs) | |
File "<eval_with_key>.5", line 4, in forward | |
def forward(self, x : torch.Tensor) -> torch.Tensor: | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 893, in forward | |
return compiled_f( | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 885, in new_func | |
return compiled_fn(args) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 271, in g | |
return f(*args) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 513, in compiled_function | |
return CompiledFunction.apply(*remove_dupe_args(args)) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 471, in forward | |
fw_outs = call_func_with_args( | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 296, in call_func_with_args | |
out = normalize_as_list(f(args)) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 186, in run | |
return model(new_inputs) | |
File "/tmp/torchinductor_ezyang/pc/cpcvl67kxf23dyi4iyfxqfpdajqgxrwhujrc6hybabfsmffl3oyq.py", line 3492, in call | |
return (buf1070, orig_primals_1, orig_primals_2, orig_primals_4, orig_primals_5, orig_primals_7, orig_primals_8, orig_primals_10, orig_primals_11, orig_primals_13, orig_primals_14, orig_primals_16, orig_primals_17, orig_primals_19, orig_primals_20, orig_primals_22, orig_primals_23, orig_primals_25, orig_primals_26, orig_primals_28, orig_primals_29, orig_primals_31, orig_primals_32, orig_primals_34, orig_primals_35, orig_primals_37, orig_primals_38, orig_primals_40, orig_primals_41, orig_primals_43, orig_primals_44, orig_primals_46, orig_primals_47, orig_primals_49, orig_primals_50, orig_primals_52, orig_primals_53, orig_primals_55, orig_primals_56, orig_primals_58, orig_primals_59, orig_primals_61, orig_primals_62, orig_primals_64, orig_primals_65, orig_primals_67, orig_primals_68, orig_primals_70, orig_primals_71, orig_primals_73, orig_primals_74, orig_primals_76, orig_primals_77, orig_primals_79, orig_primals_80, orig_primals_82, orig_primals_83, orig_primals_85, orig_primals_86, orig_primals_88, orig_primals_89, orig_primals_91, orig_primals_92, orig_primals_94, orig_primals_95, orig_primals_97, orig_primals_98, orig_primals_100, orig_primals_101, orig_primals_103, orig_primals_104, orig_primals_106, orig_primals_107, orig_primals_109, orig_primals_110, orig_primals_112, orig_primals_113, orig_primals_115, orig_primals_116, orig_primals_118, orig_primals_119, orig_primals_121, orig_primals_122, orig_primals_124, orig_primals_125, orig_primals_127, orig_primals_128, orig_primals_130, orig_primals_131, orig_primals_133, orig_primals_134, orig_primals_136, orig_primals_137, orig_primals_139, orig_primals_140, orig_primals_142, orig_primals_143, orig_primals_145, orig_primals_146, orig_primals_148, orig_primals_149, orig_primals_151, orig_primals_152, orig_primals_154, orig_primals_155, orig_primals_157, orig_primals_158, orig_primals_169, orig_primals_170, orig_primals_186, orig_primals_187, orig_primals_203, orig_primals_204, orig_primals_220, orig_primals_221, orig_primals_237, orig_primals_238, orig_primals_261, orig_primals_262, orig_primals_278, orig_primals_279, orig_primals_295, orig_primals_296, orig_primals_319, orig_primals_320, orig_primals_336, orig_primals_337, orig_primals_353, orig_primals_354, orig_primals_377, orig_primals_378, orig_primals_394, orig_primals_395, orig_primals_411, orig_primals_412, orig_primals_428, orig_primals_429, orig_primals_452, orig_primals_453, orig_primals_469, orig_primals_470, orig_primals_486, orig_primals_487, orig_primals_510, orig_primals_511, orig_primals_527, orig_primals_528, orig_primals_544, orig_primals_545, orig_primals_568, orig_primals_569, orig_primals_585, orig_primals_586, orig_primals_602, orig_primals_603, orig_primals_626, orig_primals_627, orig_primals_643, orig_primals_644, orig_primals_660, orig_primals_661, orig_primals_677, orig_primals_678, orig_primals_701, orig_primals_702, orig_primals_718, orig_primals_719, orig_primals_735, orig_primals_736, orig_primals_759, orig_primals_760, orig_primals_776, orig_primals_777, orig_primals_793, orig_primals_794, orig_primals_817, orig_primals_818, orig_primals_834, orig_primals_835, orig_primals_851, orig_primals_852, orig_primals_875, orig_primals_876, orig_primals_892, orig_primals_893, orig_primals_909, orig_primals_910, orig_primals_933, orig_primals_934, orig_primals_950, orig_primals_951, orig_primals_967, orig_primals_968, orig_primals_991, orig_primals_992, orig_primals_1008, orig_primals_1009, orig_primals_1025, orig_primals_1026, orig_primals_1042, orig_primals_1043, orig_primals_1066, orig_primals_1067, orig_primals_1083, orig_primals_1084, orig_primals_1100, orig_primals_1101, orig_primals_1124, orig_primals_1125, orig_primals_1141, orig_primals_1142, orig_primals_1158, orig_primals_1159, buf1, buf9, buf10, buf1076, buf18, buf19, buf25, buf27, buf28, buf35, buf36, buf1077, buf44, buf45, buf52, buf53, buf1078, buf61, buf62, buf69, buf70, buf1079, buf79, buf86, buf87, buf1080, buf96, buf103, buf104, buf111, buf112, buf1081, buf120, buf121, buf128, buf129, buf1082, buf137, buf138, buf145, buf146, buf1083, buf155, buf162, buf163, buf170, buf171, buf1084, buf179, buf180, buf187, buf188, buf1085, buf196, buf197, buf204, buf205, buf1086, buf214, buf221, buf222, buf229, buf230, buf1087, buf238, buf239, buf246, buf247, buf1088, buf255, buf256, buf263, buf264, buf1089, buf273, buf280, buf281, buf1090, buf290, buf297, buf298, buf305, buf306, buf1091, buf314, buf315, buf322, buf323, buf1092, buf331, buf332, buf339, buf340, buf1093, buf349, buf356, buf357, buf364, buf365, buf1094, buf373, buf374, buf381, buf382, buf1095, buf390, buf391, buf398, buf399, buf1096, buf408, buf415, buf416, buf423, buf424, buf1097, buf432, buf433, buf440, buf441, buf1098, buf449, buf450, buf457, buf458, buf1099, buf467, buf474, buf475, buf482, buf483, buf1100, buf491, buf492, buf499, buf500, buf1101, buf508, buf509, buf516, buf517, buf1102, buf526, buf533, buf534, buf1103, buf543, buf550, buf551, buf558, buf559, buf1104, buf567, buf568, buf575, buf576, buf1105, buf584, buf585, buf592, buf593, buf1106, buf602, buf609, buf610, buf617, buf618, buf1107, buf626, buf627, buf634, buf635, buf1108, buf643, buf644, buf651, buf652, buf1109, buf661, buf668, buf669, buf676, buf677, buf1110, buf685, buf686, buf693, buf694, buf1111, buf702, buf703, buf710, buf711, buf1112, buf720, buf727, buf728, buf735, buf736, buf1113, buf744, buf745, buf752, buf753, buf1114, buf761, buf762, buf769, buf770, buf1115, buf779, buf786, buf787, buf794, buf795, buf1116, buf803, buf804, buf811, buf812, buf1117, buf820, buf821, buf828, buf829, buf1118, buf838, buf845, buf846, buf853, buf854, buf1119, buf862, buf863, buf870, buf871, buf1120, buf879, buf880, buf887, buf888, buf1121, buf897, buf904, buf905, buf1122, buf914, buf921, buf922, buf929, buf930, buf1123, buf938, buf939, buf946, buf947, buf1124, buf955, buf956, buf963, buf964, buf1125, buf973, buf980, buf981, buf988, buf989, buf1126, buf997, buf998, buf1005, buf1006, buf1127, buf1014, buf1015, buf1022, buf1023, buf1128, buf1032, buf1040, buf1049, buf1055, buf1056, buf1063, buf1071, as_strided(buf1062, (1000, 2048), (2048, 1)), buf1129, buf1130, buf1131, buf1132, buf1133, buf1134, buf1135, buf1136, buf1137, buf1138, buf1139, buf1140, buf1141, buf1142, buf1143, buf1144, buf1145, buf1146, buf1147, buf1148, buf1149, buf1150, buf1151, buf1152, buf1153, buf1154, buf1155, buf1156, buf1157, buf1158, buf1159, buf1160, buf1161, buf1162, buf1163, buf1164, buf1165, buf1166, buf1167, buf1168, buf1169, buf1170, buf1171, buf1172, buf1173, buf1174, buf1175, buf1176, buf1177, s0, ) | |
NameError: name 's0' is not defined | |
TorchDynamo optimized model failed to run because of following error | |
cuda train resnet50_quantized_qat FAIL | |
Running torchbench.py resnext50_32x4d... | |
ERROR:common:'int' object has no attribute 'size' | |
While executing %sym_size : [#users=2] = placeholder[target=sym_size] | |
Original traceback: | |
Module stack: {} | |
File "/data/users/ezyang/vision/torchvision/models/resnet.py", line 279, in _forward_impl | |
x = torch.flatten(x, 1) | |
| File "/data/users/ezyang/vision/torchvision/models/resnet.py", line 285, in forward | |
return self._forward_impl(x) | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1041, in check_accuracy | |
new_result = optimized_model_iter_fn(model, example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 945, in run_n_iterations | |
self.model_iter_fn(mod, inputs, collect_outputs=False) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 332, in forward_and_backward_pass | |
cloned_inputs = clone_inputs(inputs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 337, in <graph break in forward_and_backward_pass> | |
self.grad_scaler.scale(loss).backward() | |
File "/data/users/ezyang/pytorch-tmp/torch/_tensor.py", line 488, in backward | |
torch.autograd.backward( | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/__init__.py", line 197, in backward | |
Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/function.py", line 270, in apply | |
return user_fn(self, *args) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 499, in backward | |
CompiledFunction.compiled_bw = aot_config.bw_compiler( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/optimizations/backends.py", line 555, in _wrapped_bw_compiler | |
return disable(bw_compiler(*args, **kwargs)) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 362, in bw_compiler | |
return compile_fx_inner( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/debug_utils.py", line 466, in debug_wrapper | |
compiled_fn = compiler_fn(gm, example_inputs, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/debug.py", line 177, in inner | |
return fn(*args, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/contextlib.py", line 79, in inner | |
return func(*args, **kwds) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 122, in compile_fx_inner | |
graph.run(*example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 129, in run | |
return super().run(*args) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 130, in run | |
self.env[node] = self.run_node(node) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 299, in run_node | |
result = super().run_node(n) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 171, in run_node | |
return getattr(self, n.op)(n.target, args, kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 199, in placeholder | |
sizes, strides = self.static_sizes_strides(example) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 53, in static_sizes_strides | |
size = [sympy.Integer(i) for i in ex.size()] | |
AttributeError: 'int' object has no attribute 'size' | |
While executing %sym_size : [#users=2] = placeholder[target=sym_size] | |
Original traceback: | |
Module stack: {} | |
File "/data/users/ezyang/vision/torchvision/models/resnet.py", line 279, in _forward_impl | |
x = torch.flatten(x, 1) | |
| File "/data/users/ezyang/vision/torchvision/models/resnet.py", line 285, in forward | |
return self._forward_impl(x) | |
TorchDynamo optimized model failed to run because of following error | |
cuda train resnext50_32x4d FAIL | |
Running torchbench.py shufflenet_v2_x1_0... | |
ERROR:common:'int' object has no attribute 'size' | |
While executing %sym_size : [#users=33] = placeholder[target=sym_size] | |
Original traceback: | |
Module stack: {'self_stage2': <class 'torch.nn.modules.container.Sequential'>, 'self_stage2_0': <class 'torchvision.models.shufflenetv2.InvertedResidual'>} | |
File "/data/users/ezyang/vision/torchvision/models/shufflenetv2.py", line 33, in channel_shuffle | |
x = x.view(batchsize, groups, channels_per_group, height, width) | |
| File "/data/users/ezyang/vision/torchvision/models/shufflenetv2.py", line 99, in forward | |
out = channel_shuffle(out, 2) | |
| File "/data/users/ezyang/vision/torchvision/models/shufflenetv2.py", line 157, in _forward_impl | |
x = self.stage2(x) | |
| File "/data/users/ezyang/vision/torchvision/models/shufflenetv2.py", line 166, in forward | |
return self._forward_impl(x) | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 1041, in check_accuracy | |
new_result = optimized_model_iter_fn(model, example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/eval_frame.py", line 157, in _fn | |
return fn(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/common.py", line 945, in run_n_iterations | |
self.model_iter_fn(mod, inputs, collect_outputs=False) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 332, in forward_and_backward_pass | |
cloned_inputs = clone_inputs(inputs) | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 337, in <graph break in forward_and_backward_pass> | |
self.grad_scaler.scale(loss).backward() | |
File "/data/users/ezyang/pytorch-tmp/torch/_tensor.py", line 488, in backward | |
torch.autograd.backward( | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/__init__.py", line 197, in backward | |
Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass | |
File "/data/users/ezyang/pytorch-tmp/torch/autograd/function.py", line 270, in apply | |
return user_fn(self, *args) | |
File "/data/users/ezyang/pytorch-tmp/functorch/_src/aot_autograd.py", line 499, in backward | |
CompiledFunction.compiled_bw = aot_config.bw_compiler( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/optimizations/backends.py", line 555, in _wrapped_bw_compiler | |
return disable(bw_compiler(*args, **kwargs)) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 362, in bw_compiler | |
return compile_fx_inner( | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/debug_utils.py", line 466, in debug_wrapper | |
compiled_fn = compiler_fn(gm, example_inputs, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/debug.py", line 177, in inner | |
return fn(*args, **kwargs) | |
File "/home/ezyang/local/pytorch-tmp-env/lib/python3.9/contextlib.py", line 79, in inner | |
return func(*args, **kwds) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/compile_fx.py", line 122, in compile_fx_inner | |
graph.run(*example_inputs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_dynamo/utils.py", line 87, in time_wrapper | |
r = func(*args, **kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 129, in run | |
return super().run(*args) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 130, in run | |
self.env[node] = self.run_node(node) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 299, in run_node | |
result = super().run_node(n) | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/interpreter.py", line 171, in run_node | |
return getattr(self, n.op)(n.target, args, kwargs) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 199, in placeholder | |
sizes, strides = self.static_sizes_strides(example) | |
File "/data/users/ezyang/pytorch-tmp/torch/_inductor/graph.py", line 53, in static_sizes_strides | |
size = [sympy.Integer(i) for i in ex.size()] | |
AttributeError: 'int' object has no attribute 'size' | |
While executing %sym_size : [#users=33] = placeholder[target=sym_size] | |
Original traceback: | |
Module stack: {'self_stage2': <class 'torch.nn.modules.container.Sequential'>, 'self_stage2_0': <class 'torchvision.models.shufflenetv2.InvertedResidual'>} | |
File "/data/users/ezyang/vision/torchvision/models/shufflenetv2.py", line 33, in channel_shuffle | |
x = x.view(batchsize, groups, channels_per_group, height, width) | |
| File "/data/users/ezyang/vision/torchvision/models/shufflenetv2.py", line 99, in forward | |
out = channel_shuffle(out, 2) | |
| File "/data/users/ezyang/vision/torchvision/models/shufflenetv2.py", line 157, in _forward_impl | |
x = self.stage2(x) | |
| File "/data/users/ezyang/vision/torchvision/models/shufflenetv2.py", line 166, in forward | |
return self._forward_impl(x) | |
TorchDynamo optimized model failed to run because of following error | |
cuda train shufflenet_v2_x1_0 FAIL | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 11, in <module> | |
import torch | |
File "/data/users/ezyang/pytorch-tmp/torch/__init__.py", line 1027, in <module> | |
from . import _meta_registrations | |
File "/data/users/ezyang/pytorch-tmp/torch/_meta_registrations.py", line 7, in <module> | |
from torch._decomp import _add_op_to_registry, global_decomposition_table, meta_table | |
File "/data/users/ezyang/pytorch-tmp/torch/_decomp/__init__.py", line 168, in <module> | |
import torch._decomp.decompositions | |
File "/data/users/ezyang/pytorch-tmp/torch/_decomp/decompositions.py", line 16, in <module> | |
from torch.fx.experimental.symbolic_shapes import ( | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/experimental/symbolic_shapes.py", line 12, in <module> | |
from torch._subclasses.meta_utils import MetaConverter | |
File "/data/users/ezyang/pytorch-tmp/torch/_subclasses/__init__.py", line 3, in <module> | |
from torch._subclasses.fake_tensor import ( | |
File "/data/users/ezyang/pytorch-tmp/torch/_subclasses/fake_tensor.py", line 893 | |
<<<<<<< HEAD | |
^ | |
SyntaxError: invalid syntax | |
cuda train soft_actor_critic FAIL | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 11, in <module> | |
import torch | |
File "/data/users/ezyang/pytorch-tmp/torch/__init__.py", line 1027, in <module> | |
from . import _meta_registrations | |
File "/data/users/ezyang/pytorch-tmp/torch/_meta_registrations.py", line 7, in <module> | |
from torch._decomp import _add_op_to_registry, global_decomposition_table, meta_table | |
File "/data/users/ezyang/pytorch-tmp/torch/_decomp/__init__.py", line 168, in <module> | |
import torch._decomp.decompositions | |
File "/data/users/ezyang/pytorch-tmp/torch/_decomp/decompositions.py", line 16, in <module> | |
from torch.fx.experimental.symbolic_shapes import ( | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/experimental/symbolic_shapes.py", line 12, in <module> | |
from torch._subclasses.meta_utils import MetaConverter | |
File "/data/users/ezyang/pytorch-tmp/torch/_subclasses/__init__.py", line 3, in <module> | |
from torch._subclasses.fake_tensor import ( | |
File "/data/users/ezyang/pytorch-tmp/torch/_subclasses/fake_tensor.py", line 893 | |
<<<<<<< HEAD | |
^ | |
SyntaxError: invalid syntax | |
cuda train speech_transformer FAIL | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 11, in <module> | |
import torch | |
File "/data/users/ezyang/pytorch-tmp/torch/__init__.py", line 1027, in <module> | |
from . import _meta_registrations | |
File "/data/users/ezyang/pytorch-tmp/torch/_meta_registrations.py", line 7, in <module> | |
from torch._decomp import _add_op_to_registry, global_decomposition_table, meta_table | |
File "/data/users/ezyang/pytorch-tmp/torch/_decomp/__init__.py", line 168, in <module> | |
import torch._decomp.decompositions | |
File "/data/users/ezyang/pytorch-tmp/torch/_decomp/decompositions.py", line 16, in <module> | |
from torch.fx.experimental.symbolic_shapes import ( | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/experimental/symbolic_shapes.py", line 12, in <module> | |
from torch._subclasses.meta_utils import MetaConverter | |
File "/data/users/ezyang/pytorch-tmp/torch/_subclasses/__init__.py", line 3, in <module> | |
from torch._subclasses.fake_tensor import ( | |
File "/data/users/ezyang/pytorch-tmp/torch/_subclasses/fake_tensor.py", line 893 | |
<<<<<<< HEAD | |
^ | |
SyntaxError: invalid syntax | |
cuda train squeezenet1_1 FAIL | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 11, in <module> | |
import torch | |
File "/data/users/ezyang/pytorch-tmp/torch/__init__.py", line 1027, in <module> | |
from . import _meta_registrations | |
File "/data/users/ezyang/pytorch-tmp/torch/_meta_registrations.py", line 7, in <module> | |
from torch._decomp import _add_op_to_registry, global_decomposition_table, meta_table | |
File "/data/users/ezyang/pytorch-tmp/torch/_decomp/__init__.py", line 168, in <module> | |
import torch._decomp.decompositions | |
File "/data/users/ezyang/pytorch-tmp/torch/_decomp/decompositions.py", line 16, in <module> | |
from torch.fx.experimental.symbolic_shapes import ( | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/experimental/symbolic_shapes.py", line 12, in <module> | |
from torch._subclasses.meta_utils import MetaConverter | |
File "/data/users/ezyang/pytorch-tmp/torch/_subclasses/__init__.py", line 3, in <module> | |
from torch._subclasses.fake_tensor import ( | |
File "/data/users/ezyang/pytorch-tmp/torch/_subclasses/fake_tensor.py", line 893 | |
<<<<<<< HEAD | |
^ | |
SyntaxError: invalid syntax | |
cuda train tacotron2 FAIL | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 11, in <module> | |
import torch | |
File "/data/users/ezyang/pytorch-tmp/torch/__init__.py", line 1027, in <module> | |
from . import _meta_registrations | |
File "/data/users/ezyang/pytorch-tmp/torch/_meta_registrations.py", line 7, in <module> | |
from torch._decomp import _add_op_to_registry, global_decomposition_table, meta_table | |
File "/data/users/ezyang/pytorch-tmp/torch/_decomp/__init__.py", line 168, in <module> | |
import torch._decomp.decompositions | |
File "/data/users/ezyang/pytorch-tmp/torch/_decomp/decompositions.py", line 16, in <module> | |
from torch.fx.experimental.symbolic_shapes import ( | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/experimental/symbolic_shapes.py", line 12, in <module> | |
from torch._subclasses.meta_utils import MetaConverter | |
File "/data/users/ezyang/pytorch-tmp/torch/_subclasses/__init__.py", line 3, in <module> | |
from torch._subclasses.fake_tensor import ( | |
File "/data/users/ezyang/pytorch-tmp/torch/_subclasses/fake_tensor.py", line 893 | |
<<<<<<< HEAD | |
^ | |
SyntaxError: invalid syntax | |
cuda train timm_efficientdet FAIL | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 11, in <module> | |
import torch | |
File "/data/users/ezyang/pytorch-tmp/torch/__init__.py", line 1027, in <module> | |
from . import _meta_registrations | |
File "/data/users/ezyang/pytorch-tmp/torch/_meta_registrations.py", line 7, in <module> | |
from torch._decomp import _add_op_to_registry, global_decomposition_table, meta_table | |
File "/data/users/ezyang/pytorch-tmp/torch/_decomp/__init__.py", line 168, in <module> | |
import torch._decomp.decompositions | |
File "/data/users/ezyang/pytorch-tmp/torch/_decomp/decompositions.py", line 16, in <module> | |
from torch.fx.experimental.symbolic_shapes import ( | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/experimental/symbolic_shapes.py", line 12, in <module> | |
from torch._subclasses.meta_utils import MetaConverter | |
File "/data/users/ezyang/pytorch-tmp/torch/_subclasses/__init__.py", line 3, in <module> | |
from torch._subclasses.fake_tensor import ( | |
File "/data/users/ezyang/pytorch-tmp/torch/_subclasses/fake_tensor.py", line 893 | |
<<<<<<< HEAD | |
^ | |
SyntaxError: invalid syntax | |
cuda train timm_efficientnet FAIL | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 11, in <module> | |
import torch | |
File "/data/users/ezyang/pytorch-tmp/torch/__init__.py", line 1027, in <module> | |
from . import _meta_registrations | |
File "/data/users/ezyang/pytorch-tmp/torch/_meta_registrations.py", line 7, in <module> | |
from torch._decomp import _add_op_to_registry, global_decomposition_table, meta_table | |
File "/data/users/ezyang/pytorch-tmp/torch/_decomp/__init__.py", line 168, in <module> | |
import torch._decomp.decompositions | |
File "/data/users/ezyang/pytorch-tmp/torch/_decomp/decompositions.py", line 16, in <module> | |
from torch.fx.experimental.symbolic_shapes import ( | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/experimental/symbolic_shapes.py", line 12, in <module> | |
from torch._subclasses.meta_utils import MetaConverter | |
File "/data/users/ezyang/pytorch-tmp/torch/_subclasses/__init__.py", line 3, in <module> | |
from torch._subclasses.fake_tensor import ( | |
File "/data/users/ezyang/pytorch-tmp/torch/_subclasses/fake_tensor.py", line 893 | |
<<<<<<< HEAD | |
^ | |
SyntaxError: invalid syntax | |
cuda train timm_regnet FAIL | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 11, in <module> | |
import torch | |
File "/data/users/ezyang/pytorch-tmp/torch/__init__.py", line 1027, in <module> | |
from . import _meta_registrations | |
File "/data/users/ezyang/pytorch-tmp/torch/_meta_registrations.py", line 7, in <module> | |
from torch._decomp import _add_op_to_registry, global_decomposition_table, meta_table | |
File "/data/users/ezyang/pytorch-tmp/torch/_decomp/__init__.py", line 168, in <module> | |
import torch._decomp.decompositions | |
File "/data/users/ezyang/pytorch-tmp/torch/_decomp/decompositions.py", line 16, in <module> | |
from torch.fx.experimental.symbolic_shapes import ( | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/experimental/symbolic_shapes.py", line 12, in <module> | |
from torch._subclasses.meta_utils import MetaConverter | |
File "/data/users/ezyang/pytorch-tmp/torch/_subclasses/__init__.py", line 3, in <module> | |
from torch._subclasses.fake_tensor import ( | |
File "/data/users/ezyang/pytorch-tmp/torch/_subclasses/fake_tensor.py", line 893 | |
<<<<<<< HEAD | |
^ | |
SyntaxError: invalid syntax | |
cuda train timm_resnest FAIL | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 11, in <module> | |
import torch | |
File "/data/users/ezyang/pytorch-tmp/torch/__init__.py", line 1027, in <module> | |
from . import _meta_registrations | |
File "/data/users/ezyang/pytorch-tmp/torch/_meta_registrations.py", line 7, in <module> | |
from torch._decomp import _add_op_to_registry, global_decomposition_table, meta_table | |
File "/data/users/ezyang/pytorch-tmp/torch/_decomp/__init__.py", line 168, in <module> | |
import torch._decomp.decompositions | |
File "/data/users/ezyang/pytorch-tmp/torch/_decomp/decompositions.py", line 16, in <module> | |
from torch.fx.experimental.symbolic_shapes import ( | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/experimental/symbolic_shapes.py", line 12, in <module> | |
from torch._subclasses.meta_utils import MetaConverter | |
File "/data/users/ezyang/pytorch-tmp/torch/_subclasses/__init__.py", line 3, in <module> | |
from torch._subclasses.fake_tensor import ( | |
File "/data/users/ezyang/pytorch-tmp/torch/_subclasses/fake_tensor.py", line 893 | |
<<<<<<< HEAD | |
^ | |
SyntaxError: invalid syntax | |
cuda train timm_vision_transformer FAIL | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 11, in <module> | |
import torch | |
File "/data/users/ezyang/pytorch-tmp/torch/__init__.py", line 1027, in <module> | |
from . import _meta_registrations | |
File "/data/users/ezyang/pytorch-tmp/torch/_meta_registrations.py", line 7, in <module> | |
from torch._decomp import _add_op_to_registry, global_decomposition_table, meta_table | |
File "/data/users/ezyang/pytorch-tmp/torch/_decomp/__init__.py", line 168, in <module> | |
import torch._decomp.decompositions | |
File "/data/users/ezyang/pytorch-tmp/torch/_decomp/decompositions.py", line 16, in <module> | |
from torch.fx.experimental.symbolic_shapes import ( | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/experimental/symbolic_shapes.py", line 12, in <module> | |
from torch._subclasses.meta_utils import MetaConverter | |
File "/data/users/ezyang/pytorch-tmp/torch/_subclasses/__init__.py", line 3, in <module> | |
from torch._subclasses.fake_tensor import ( | |
File "/data/users/ezyang/pytorch-tmp/torch/_subclasses/fake_tensor.py", line 893 | |
<<<<<<< HEAD | |
^ | |
SyntaxError: invalid syntax | |
cuda train timm_vision_transformer_large FAIL | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 11, in <module> | |
import torch | |
File "/data/users/ezyang/pytorch-tmp/torch/__init__.py", line 1027, in <module> | |
from . import _meta_registrations | |
File "/data/users/ezyang/pytorch-tmp/torch/_meta_registrations.py", line 7, in <module> | |
from torch._decomp import _add_op_to_registry, global_decomposition_table, meta_table | |
File "/data/users/ezyang/pytorch-tmp/torch/_decomp/__init__.py", line 168, in <module> | |
import torch._decomp.decompositions | |
File "/data/users/ezyang/pytorch-tmp/torch/_decomp/decompositions.py", line 16, in <module> | |
from torch.fx.experimental.symbolic_shapes import ( | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/experimental/symbolic_shapes.py", line 12, in <module> | |
from torch._subclasses.meta_utils import MetaConverter | |
File "/data/users/ezyang/pytorch-tmp/torch/_subclasses/__init__.py", line 3, in <module> | |
from torch._subclasses.fake_tensor import ( | |
File "/data/users/ezyang/pytorch-tmp/torch/_subclasses/fake_tensor.py", line 893 | |
<<<<<<< HEAD | |
^ | |
SyntaxError: invalid syntax | |
cuda train timm_vovnet FAIL | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 11, in <module> | |
import torch | |
File "/data/users/ezyang/pytorch-tmp/torch/__init__.py", line 1027, in <module> | |
from . import _meta_registrations | |
File "/data/users/ezyang/pytorch-tmp/torch/_meta_registrations.py", line 7, in <module> | |
from torch._decomp import _add_op_to_registry, global_decomposition_table, meta_table | |
File "/data/users/ezyang/pytorch-tmp/torch/_decomp/__init__.py", line 168, in <module> | |
import torch._decomp.decompositions | |
File "/data/users/ezyang/pytorch-tmp/torch/_decomp/decompositions.py", line 16, in <module> | |
from torch.fx.experimental.symbolic_shapes import ( | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/experimental/symbolic_shapes.py", line 12, in <module> | |
from torch._subclasses.meta_utils import MetaConverter | |
File "/data/users/ezyang/pytorch-tmp/torch/_subclasses/__init__.py", line 3, in <module> | |
from torch._subclasses.fake_tensor import ( | |
File "/data/users/ezyang/pytorch-tmp/torch/_subclasses/fake_tensor.py", line 893 | |
<<<<<<< HEAD | |
^ | |
SyntaxError: invalid syntax | |
cuda train tts_angular FAIL | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 11, in <module> | |
import torch | |
File "/data/users/ezyang/pytorch-tmp/torch/__init__.py", line 1027, in <module> | |
from . import _meta_registrations | |
File "/data/users/ezyang/pytorch-tmp/torch/_meta_registrations.py", line 7, in <module> | |
from torch._decomp import _add_op_to_registry, global_decomposition_table, meta_table | |
File "/data/users/ezyang/pytorch-tmp/torch/_decomp/__init__.py", line 168, in <module> | |
import torch._decomp.decompositions | |
File "/data/users/ezyang/pytorch-tmp/torch/_decomp/decompositions.py", line 16, in <module> | |
from torch.fx.experimental.symbolic_shapes import ( | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/experimental/symbolic_shapes.py", line 12, in <module> | |
from torch._subclasses.meta_utils import MetaConverter | |
File "/data/users/ezyang/pytorch-tmp/torch/_subclasses/__init__.py", line 3, in <module> | |
from torch._subclasses.fake_tensor import ( | |
File "/data/users/ezyang/pytorch-tmp/torch/_subclasses/fake_tensor.py", line 893 | |
<<<<<<< HEAD | |
^ | |
SyntaxError: invalid syntax | |
cuda train vgg16 FAIL | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 11, in <module> | |
import torch | |
File "/data/users/ezyang/pytorch-tmp/torch/__init__.py", line 1027, in <module> | |
from . import _meta_registrations | |
File "/data/users/ezyang/pytorch-tmp/torch/_meta_registrations.py", line 7, in <module> | |
from torch._decomp import _add_op_to_registry, global_decomposition_table, meta_table | |
File "/data/users/ezyang/pytorch-tmp/torch/_decomp/__init__.py", line 168, in <module> | |
import torch._decomp.decompositions | |
File "/data/users/ezyang/pytorch-tmp/torch/_decomp/decompositions.py", line 16, in <module> | |
from torch.fx.experimental.symbolic_shapes import ( | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/experimental/symbolic_shapes.py", line 12, in <module> | |
from torch._subclasses.meta_utils import MetaConverter | |
File "/data/users/ezyang/pytorch-tmp/torch/_subclasses/__init__.py", line 3, in <module> | |
from torch._subclasses.fake_tensor import ( | |
File "/data/users/ezyang/pytorch-tmp/torch/_subclasses/fake_tensor.py", line 893 | |
<<<<<<< HEAD | |
^ | |
SyntaxError: invalid syntax | |
cuda train vision_maskrcnn FAIL | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/torchbench.py", line 11, in <module> | |
import torch | |
File "/data/users/ezyang/pytorch-tmp/torch/__init__.py", line 1027, in <module> | |
from . import _meta_registrations | |
File "/data/users/ezyang/pytorch-tmp/torch/_meta_registrations.py", line 7, in <module> | |
from torch._decomp import _add_op_to_registry, global_decomposition_table, meta_table | |
File "/data/users/ezyang/pytorch-tmp/torch/_decomp/__init__.py", line 168, in <module> | |
import torch._decomp.decompositions | |
File "/data/users/ezyang/pytorch-tmp/torch/_decomp/decompositions.py", line 16, in <module> | |
from torch.fx.experimental.symbolic_shapes import ( | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/experimental/symbolic_shapes.py", line 12, in <module> | |
from torch._subclasses.meta_utils import MetaConverter | |
File "/data/users/ezyang/pytorch-tmp/torch/_subclasses/__init__.py", line 3, in <module> | |
from torch._subclasses.fake_tensor import ( | |
File "/data/users/ezyang/pytorch-tmp/torch/_subclasses/fake_tensor.py", line 893 | |
<<<<<<< HEAD | |
^ | |
SyntaxError: invalid syntax | |
cuda train yolov3 FAIL | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/huggingface.py", line 10, in <module> | |
import torch | |
File "/data/users/ezyang/pytorch-tmp/torch/__init__.py", line 1027, in <module> | |
from . import _meta_registrations | |
File "/data/users/ezyang/pytorch-tmp/torch/_meta_registrations.py", line 7, in <module> | |
from torch._decomp import _add_op_to_registry, global_decomposition_table, meta_table | |
File "/data/users/ezyang/pytorch-tmp/torch/_decomp/__init__.py", line 168, in <module> | |
import torch._decomp.decompositions | |
File "/data/users/ezyang/pytorch-tmp/torch/_decomp/decompositions.py", line 16, in <module> | |
from torch.fx.experimental.symbolic_shapes import ( | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/experimental/symbolic_shapes.py", line 12, in <module> | |
from torch._subclasses.meta_utils import MetaConverter | |
File "/data/users/ezyang/pytorch-tmp/torch/_subclasses/__init__.py", line 3, in <module> | |
from torch._subclasses.fake_tensor import ( | |
File "/data/users/ezyang/pytorch-tmp/torch/_subclasses/fake_tensor.py", line 893 | |
<<<<<<< HEAD | |
^ | |
SyntaxError: invalid syntax | |
Traceback (most recent call last): | |
File "/data/users/ezyang/pytorch-tmp/benchmarks/dynamo/timm_models.py", line 11, in <module> | |
import torch | |
File "/data/users/ezyang/pytorch-tmp/torch/__init__.py", line 1027, in <module> | |
from . import _meta_registrations | |
File "/data/users/ezyang/pytorch-tmp/torch/_meta_registrations.py", line 7, in <module> | |
from torch._decomp import _add_op_to_registry, global_decomposition_table, meta_table | |
File "/data/users/ezyang/pytorch-tmp/torch/_decomp/__init__.py", line 168, in <module> | |
import torch._decomp.decompositions | |
File "/data/users/ezyang/pytorch-tmp/torch/_decomp/decompositions.py", line 16, in <module> | |
from torch.fx.experimental.symbolic_shapes import ( | |
File "/data/users/ezyang/pytorch-tmp/torch/fx/experimental/symbolic_shapes.py", line 12, in <module> | |
from torch._subclasses.meta_utils import MetaConverter | |
File "/data/users/ezyang/pytorch-tmp/torch/_subclasses/__init__.py", line 3, in <module> | |
from torch._subclasses.fake_tensor import ( | |
File "/data/users/ezyang/pytorch-tmp/torch/_subclasses/fake_tensor.py", line 893 | |
<<<<<<< HEAD | |
^ | |
SyntaxError: invalid syntax |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment