Last active
December 3, 2021 10:36
-
-
Save robin-p-schmitt/d59570c322891fcbb49bc0a7d6944562 to your computer and use it in GitHub Desktop.
ReinterpretDataLayer: beam information of new DimensionTag is not set correctly
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
| layer <network via test_concat_new_dim_tag>/'data' output: Data{'data', [B,T|'time'[B],F|F'feature:data'(5)]} | |
| layer <network via test_concat_new_dim_tag>/'data_new' output: Data{'data_new_output', [B,T|'new-time'[?],F|F'feature:data'(5)]} | |
| layer <network via test_concat_new_dim_tag>/'output' output: Data{'output_output', [T|'time'[B&Beam{'output/output'}(3)],B&Beam{'output/output'}(3)], dtype='int32', sparse_dim=DimensionTag{F'classes:sparse-dim'(5)}} | |
| Rec layer 'output' (search True, train False) sub net: | |
| Input layers moved out of loop: (#: 0) | |
| None | |
| Output layers moved out of loop: (#: 0) | |
| None | |
| Layers in loop: (#: 7) | |
| output | |
| output_prob | |
| data_red | |
| data_concat | |
| prev_out | |
| prev_out1 | |
| prev_out0 | |
| Unused layers: (#: 0) | |
| None | |
| layer <network via test_concat_new_dim_tag>/output(rec-subnet)/'prev_out0' output: Data{'prev_out0_output', [B&Beam{'output/prev:output'}(3)], dtype='int32', ctx=loop('time'[B])} | |
| layer <network via test_concat_new_dim_tag>/output(rec-subnet)/'prev_out1' output: Data{'prev_out1_output', [B&Beam{'output/prev:output'}(3)], ctx=loop('time'[B])} | |
| layer <network via test_concat_new_dim_tag>/output(rec-subnet)/'prev_out' output: Data{'prev_out_output', [B&Beam{'output/prev:output'}(3),F|'prev_out_expand_dims'(1)], ctx=loop('time'[B])} | |
| layer <network via test_concat_new_dim_tag>/output(rec-subnet)/'data_concat' output: Data{'data_concat_output', [B&Beam{'output/prev:output'}(3),T|'new-time'[B&Beam{'output/prev:output'}(3)],F|F'data_concat_output_feature'(6)], ctx=loop('time'[B])} | |
| Exception creating layer <network via test_concat_new_dim_tag>/output(rec-subnet)/'data_concat' of class CopyLayer with opts: | |
| {'_name': 'data_concat', | |
| '_network': <TFNetwork '<network via test_concat_new_dim_tag>/output(rec-subnet)' parent_layer=<RecLayer 'output' out_type=Data{[T|'time'[B&Beam{'output/output'}(3)],B&Beam{'output/output'}(3)], dtype='int32', sparse_dim=DimensionTag{F'classes:sparse-dim'(5)}}> train=False search>, | |
| '_src_common_search_choices': <SearchChoices owner='prev:output' beam_size=3 beam_scores=shaped:(None,None)>, | |
| 'name': 'data_concat', | |
| 'network': <TFNetwork '<network via test_concat_new_dim_tag>/output(rec-subnet)' parent_layer=<RecLayer 'output' out_type=Data{[T|'time'[B&Beam{'output/output'}(3)],B&Beam{'output/output'}(3)], dtype='int32', sparse_dim=DimensionTag{F'classes:sparse-dim'(5)}}> train=False search>, | |
| 'output': Data{'data_concat_output', [B&Beam{'output/prev:output'}(3),T|'new-time'[B&Beam{'output/prev:output'}(3)],F|F'data_concat_output_feature'(6)], ctx=loop('time'[B])}, | |
| 'sources': [<SelectSearchSourcesLayer 'data_new' <SearchChoices owner='prev:output' beam_size=3 beam_scores=shaped:(None,None)> out_type=Data{[B&Beam{'output/prev:output'}(3),T|'new-time'[B&Beam{'output/prev:output'}(3)],F|F'feature:data'(5)]}>, | |
| <ExpandDimsLayer output/'prev_out' out_type=Data{[B&Beam{'output/prev:output'}(3),F|'prev_out_expand_dims'(1)], ctx=loop('time'[B])}>]} | |
| Exception creating layer <network via test_concat_new_dim_tag>/'output' of class RecLayer with opts: | |
| {'_name': 'output', | |
| '_network': <TFNetwork '<network via test_concat_new_dim_tag>' train=False search>, | |
| 'axis': DimensionTag{'time'[B]}, | |
| 'n_out': <class 'returnn.util.basic.NotSpecified'>, | |
| 'name': 'output', | |
| 'network': <TFNetwork '<network via test_concat_new_dim_tag>' train=False search>, | |
| 'output': Data{'output_output', [T|'time'[B&Beam{'output/output'}(3)],B&Beam{'output/output'}(3)], dtype='int32', sparse_dim=DimensionTag{F'classes:sparse-dim'(5)}}, | |
| 'sources': [<SourceLayer 'data' out_type=Data{[B,T|'time'[B],F|F'feature:data'(5)]}>], | |
| 'unit': <_SubnetworkRecCell '<network via test_concat_new_dim_tag>/output(rec-subnet)'>} | |
| ERROR: Got exception during in-loop construction of layer 'data_concat': | |
| AssertionError: | |
| Template network (check out types / shapes): | |
| output: <_TemplateLayer(ChoiceLayer)(:template:choice) output/'output' out_type=Data{[B&Beam{'output/output'}(3)], dtype='int32', sparse_dim=DimensionTag{F'classes:sparse-dim'(5)}, ctx=loop('time'[B])} (construction stack None)> | |
| output_prob: <_TemplateLayer(SoftmaxLayer)(:template:softmax) output/'output_prob' out_type=Data{[B&Beam{'output/prev:output'}(3),F|F'output_prob:feature-dense'(5)], ctx=loop('time'[B])} (construction stack 'output')> | |
| data_red: <_TemplateLayer(ReduceLayer)(:template:reduce) output/'data_red' out_type=Data{[B&Beam{'output/prev:output'}(3),F|F'data_concat_output_feature'(6)], ctx=loop('time'[B])} (construction stack 'output_prob')> | |
| data_concat: <_TemplateLayer(CopyLayer)(:template:copy) output/'data_concat' out_type=Data{[B&Beam{'output/prev:output'}(3),T|'new-time'[B&Beam{'output/prev:output'}(3)],F|F'data_concat_output_feature'(6)], ctx=loop('time'[B])} (construction stack 'data_red')> | |
| prev_out: <_TemplateLayer(ExpandDimsLayer)(:template:expand_dims) output/'prev_out' out_type=Data{[B&Beam{'output/prev:output'}(3),F|'prev_out_expand_dims'(1)], ctx=loop('time'[B])} (construction stack 'data_concat')> | |
| prev_out1: <_TemplateLayer(CastLayer)(:template:cast) output/'prev_out1' out_type=Data{[B&Beam{'output/prev:output'}(3)], ctx=loop('time'[B])} (construction stack 'prev_out')> | |
| prev_out0: <_TemplateLayer(ReinterpretDataLayer)(:template:reinterpret_data) output/'prev_out0' out_type=Data{[B&Beam{'output/prev:output'}(3)], dtype='int32', ctx=loop('time'[B])} (construction stack 'prev_out1')> | |
| EXCEPTION | |
| Traceback (most recent call last): | |
| File "/home/robin/.local/share/JetBrains/Toolbox/apps/PyCharm-P/ch-0/212.4746.96/plugins/python/helpers/pycharm/_jb_unittest_runner.py", line 35, in <module> | |
| line: sys.exit(main(argv=args, module=None, testRunner=unittestpy.TeamcityTestRunner, buffer=not JB_DISABLE_BUFFERING)) | |
| locals: | |
| sys = <local> <module 'sys' (built-in)> | |
| sys.exit = <local> <built-in function exit> | |
| main = <local> <class 'unittest.main.TestProgram'> | |
| argv = <not found> | |
| args = <local> ['python -m unittest', 'test_TFNetworkLayer.test_concat_new_dim_tag'], _[0]: {len = 18} | |
| module = <not found> | |
| testRunner = <not found> | |
| unittestpy = <local> <module 'teamcity.unittestpy' from '/home/robin/.local/share/JetBrains/Toolbox/apps/PyCharm-P/ch-0/212.4746.96/plugins/python/helpers/pycharm/teamcity/unittestpy.py'> | |
| unittestpy.TeamcityTestRunner = <local> <class 'teamcity.unittestpy.TeamcityTestRunner'> | |
| buffer = <not found> | |
| JB_DISABLE_BUFFERING = <local> False | |
| File "/home/robin/miniconda3/envs/i6_env/lib/python3.9/unittest/main.py", line 100, in TestProgram.__init__ | |
| line: self.parseArgs(argv) | |
| locals: | |
| self = <local> <unittest.main.TestProgram object at 0x7f9608665610> | |
| self.parseArgs = <local> <bound method TestProgram.parseArgs of <unittest.main.TestProgram object at 0x7f9608665610>> | |
| argv = <local> ['python -m unittest', 'test_TFNetworkLayer.test_concat_new_dim_tag'], _[0]: {len = 18} | |
| File "/home/robin/miniconda3/envs/i6_env/lib/python3.9/unittest/main.py", line 147, in TestProgram.parseArgs | |
| line: self.createTests() | |
| locals: | |
| self = <local> <unittest.main.TestProgram object at 0x7f9608665610> | |
| self.createTests = <local> <bound method TestProgram.createTests of <unittest.main.TestProgram object at 0x7f9608665610>> | |
| File "/home/robin/miniconda3/envs/i6_env/lib/python3.9/unittest/main.py", line 158, in TestProgram.createTests | |
| line: self.test = self.testLoader.loadTestsFromNames(self.testNames, | |
| self.module) | |
| locals: | |
| self = <local> <unittest.main.TestProgram object at 0x7f9608665610> | |
| self.test = <local> !AttributeError: 'TestProgram' object has no attribute 'test' | |
| self.testLoader = <local> <unittest.loader.TestLoader object at 0x7f96086d9040> | |
| self.testLoader.loadTestsFromNames = <local> <bound method TestLoader.loadTestsFromNames of <unittest.loader.TestLoader object at 0x7f96086d9040>> | |
| self.testNames = <local> ['test_TFNetworkLayer.test_concat_new_dim_tag'], _[0]: {len = 43} | |
| self.module = <local> None | |
| File "/home/robin/miniconda3/envs/i6_env/lib/python3.9/unittest/loader.py", line 220, in TestLoader.loadTestsFromNames | |
| line: suites = [self.loadTestsFromName(name, module) for name in names] | |
| locals: | |
| suites = <not found> | |
| self = <local> <unittest.loader.TestLoader object at 0x7f96086d9040> | |
| self.loadTestsFromName = <local> <bound method TestLoader.loadTestsFromName of <unittest.loader.TestLoader object at 0x7f96086d9040>> | |
| name = <not found> | |
| module = <local> None | |
| names = <local> ['test_TFNetworkLayer.test_concat_new_dim_tag'], _[0]: {len = 43} | |
| File "/home/robin/miniconda3/envs/i6_env/lib/python3.9/unittest/loader.py", line 220, in <listcomp> | |
| line: suites = [self.loadTestsFromName(name, module) for name in names] | |
| locals: | |
| suites = <not found> | |
| self = <local> <unittest.loader.TestLoader object at 0x7f96086d9040> | |
| self.loadTestsFromName = <local> <bound method TestLoader.loadTestsFromName of <unittest.loader.TestLoader object at 0x7f96086d9040>> | |
| name = <local> 'test_TFNetworkLayer.test_concat_new_dim_tag', len = 43 | |
| module = <local> None | |
| names = <not found> | |
| File "/home/robin/miniconda3/envs/i6_env/lib/python3.9/unittest/loader.py", line 205, in TestLoader.loadTestsFromName | |
| line: test = obj() | |
| locals: | |
| test = <not found> | |
| obj = <local> <function test_concat_new_dim_tag at 0x7f95ce77b940> | |
| File "/mnt/projects/i6/returnn/tests/test_TFNetworkLayer.py", line 398, in test_concat_new_dim_tag | |
| line: net.construct_from_dict({ | |
| "data_new": {"class": "reinterpret_data", "from": "data", | |
| "set_dim_tags": {"t": new_time_tag} | |
| }, | |
| "output": {"class": "rec", "from": "data", "unit": { | |
| "prev_out0": { | |
| "class": "reinterpret_data", "from": "prev:output", "set_sparse": False}, | |
| "prev_out1": {"class": "cast", "from": "prev_out0", "dtype": "float32"}, | |
| "prev_out": {"class": "expand_dims", "from": "prev_out1", "axis": "f"}, | |
| "data_concat": { | |
| "class": "copy", "from": ["base:data_new", "prev_out"] | |
| }, | |
| "data_red": {"class": "reduce", "from": "data_concat", "axis": "stag:new-time", "mode": "max"}, | |
| "output_prob": {"class": "softmax", "from": "data_red", "target": "classes", "loss": "ce"}, | |
| "output": { | |
| "class": "choice", "from": "output_prob", "beam_size": 3, "target": "classes", | |
| "input_type": "prob", "initial_output": 0} | |
| }} | |
| }) | |
| locals: | |
| net = <local> <TFNetwork '<network via test_concat_new_dim_tag>' train=False search> | |
| net.construct_from_dict = <local> <bound method TFNetwork.construct_from_dict of <TFNetwork '<network via test_concat_new_dim_tag>' train=False search>> | |
| new_time_tag = <local> DimensionTag{'new-time'[?]} | |
| File "/mnt/projects/i6/returnn/returnn/tf/network.py", line 609, in TFNetwork.construct_from_dict | |
| line: self.construct_layer(net_dict, name, get_layer=get_layer) | |
| locals: | |
| self = <local> <TFNetwork '<network via test_concat_new_dim_tag>' train=False search> | |
| self.construct_layer = <local> <bound method TFNetwork.construct_layer of <TFNetwork '<network via test_concat_new_dim_tag>' train=False search>> | |
| net_dict = <local> {'data_new': {'class': 'reinterpret_data', 'from': 'data', 'set_dim_tags': {'t': DimensionTag{'new-time'[?]}}}, 'output': {'class': 'rec', 'from': 'data', 'unit': {'prev_out0': {'class': 'reinterpret_data', 'from': 'prev:output', 'set_sparse': False}, 'prev_out1': {'class': 'cast', 'from': 'prev_... | |
| name = <local> 'output', len = 6 | |
| get_layer = <local> None | |
| File "/mnt/projects/i6/returnn/returnn/tf/network.py", line 935, in TFNetwork.construct_layer | |
| line: return add_layer(name=name_with_prefix, layer_class=layer_class, **layer_desc) | |
| locals: | |
| add_layer = <local> <bound method TFNetwork.add_layer of <TFNetwork '<network via test_concat_new_dim_tag>' train=False search>> | |
| name = <local> 'output', len = 6 | |
| name_with_prefix = <local> 'output', len = 6 | |
| layer_class = <local> <class 'returnn.tf.layers.rec.RecLayer'> | |
| layer_desc = <local> {'_network': <TFNetwork '<network via test_concat_new_dim_tag>' train=False search>, '_name': 'output', 'n_out': <class 'returnn.util.basic.NotSpecified'>, 'sources': [<SourceLayer 'data' out_type=Data{[B,T|'time'[B],F|F'feature:data'(5)]}>], 'axis': DimensionTag{'time'[B]}, 'unit': <_SubnetworkR..., len = 6 | |
| File "/mnt/projects/i6/returnn/returnn/tf/network.py", line 1082, in TFNetwork.add_layer | |
| line: layer = self._create_layer(name=name, layer_class=layer_class, **layer_desc) | |
| locals: | |
| layer = <not found> | |
| self = <local> <TFNetwork '<network via test_concat_new_dim_tag>' train=False search> | |
| self._create_layer = <local> <bound method TFNetwork._create_layer of <TFNetwork '<network via test_concat_new_dim_tag>' train=False search>> | |
| name = <local> 'output', len = 6 | |
| layer_class = <local> <class 'returnn.tf.layers.rec.RecLayer'> | |
| layer_desc = <local> {'_network': <TFNetwork '<network via test_concat_new_dim_tag>' train=False search>, '_name': 'output', 'n_out': <class 'returnn.util.basic.NotSpecified'>, 'sources': [<SourceLayer 'data' out_type=Data{[B,T|'time'[B],F|F'feature:data'(5)]}>], 'axis': DimensionTag{'time'[B]}, 'unit': <_SubnetworkR..., len = 6 | |
| File "/mnt/projects/i6/returnn/returnn/tf/network.py", line 1002, in TFNetwork._create_layer | |
| line: layer = layer_class(**layer_desc) | |
| locals: | |
| layer = <not found> | |
| layer_class = <local> <class 'returnn.tf.layers.rec.RecLayer'> | |
| layer_desc = <local> {'_network': <TFNetwork '<network via test_concat_new_dim_tag>' train=False search>, '_name': 'output', 'n_out': <class 'returnn.util.basic.NotSpecified'>, 'sources': [<SourceLayer 'data' out_type=Data{[B,T|'time'[B],F|F'feature:data'(5)]}>], 'axis': DimensionTag{'time'[B]}, 'unit': <_SubnetworkR..., len = 9 | |
| File "/mnt/projects/i6/returnn/returnn/tf/layers/rec.py", line 250, in RecLayer.__init__ | |
| line: y = self._get_output_subnet_unit(self.cell) | |
| locals: | |
| y = <not found> | |
| self = <local> <RecLayer 'output' out_type=Data{[T|'time'[B&Beam{'output/output'}(3)],B&Beam{'output/output'}(3)], dtype='int32', sparse_dim=DimensionTag{F'classes:sparse-dim'(5)}}> | |
| self._get_output_subnet_unit = <local> <bound method RecLayer._get_output_subnet_unit of <RecLayer 'output' out_type=Data{[T|'time'[B&Beam{'output/output'}(3)],B&Beam{'output/output'}(3)], dtype='int32', sparse_dim=DimensionTag{F'classes:sparse-dim'(5)}}>> | |
| self.cell = <local> <_SubnetworkRecCell '<network via test_concat_new_dim_tag>/output(rec-subnet)'> | |
| File "/mnt/projects/i6/returnn/returnn/tf/layers/rec.py", line 1037, in RecLayer._get_output_subnet_unit | |
| line: output = cell.get_output() | |
| locals: | |
| output = <not found> | |
| cell = <local> <_SubnetworkRecCell '<network via test_concat_new_dim_tag>/output(rec-subnet)'> | |
| cell.get_output = <local> <bound method _SubnetworkRecCell.get_output of <_SubnetworkRecCell '<network via test_concat_new_dim_tag>/output(rec-subnet)'>> | |
| File "/mnt/projects/i6/returnn/returnn/tf/layers/rec.py", line 2862, in _SubnetworkRecCell.get_output | |
| line: final_loop_vars = self._while_loop( | |
| cond=cond, | |
| body=body, | |
| loop_vars=init_loop_vars, | |
| shape_invariants=shape_invariants) | |
| locals: | |
| final_loop_vars = <not found> | |
| self = <local> <_SubnetworkRecCell '<network via test_concat_new_dim_tag>/output(rec-subnet)'> | |
| self._while_loop = <local> <bound method _SubnetworkRecCell._while_loop of <_SubnetworkRecCell '<network via test_concat_new_dim_tag>/output(rec-subnet)'>> | |
| cond = <local> <function _SubnetworkRecCell.get_output.<locals>.cond at 0x7f95ce411c10> | |
| body = <local> <function _SubnetworkRecCell.get_output.<locals>.body at 0x7f95ce411ca0> | |
| loop_vars = <not found> | |
| init_loop_vars = <local> (<tf.Tensor 'output/rec/initial_i:0' shape=() dtype=int32>, ([<tf.Tensor 'output/rec/output/init_output_const/constant_with_shape/Mul:0' shape=(?,) dtype=int32>], [[<tf.Tensor 'output/rec/output/zeros:0' shape=(?, 1) dtype=float32>, <tf.Tensor 'output/rec/output/zeros_1:0' shape=(?, 3) dtype=int3... | |
| shape_invariants = <local> (TensorShape([]), ([TensorShape([Dimension(None)])], [[TensorShape([Dimension(None), Dimension(None)]), TensorShape([Dimension(None), Dimension(None)])]]), [TensorShape(None), TensorShape(None)]), _[0]: {len = 0} | |
| File "/mnt/projects/i6/returnn/returnn/tf/layers/rec.py", line 2091, in _SubnetworkRecCell._while_loop | |
| line: return tf.while_loop( | |
| cond=cond, | |
| body=body, | |
| loop_vars=loop_vars, | |
| shape_invariants=shape_invariants, | |
| back_prop=self.parent_rec_layer.back_prop) | |
| locals: | |
| tf = <global> <module 'tensorflow' from '/home/robin/miniconda3/envs/i6_env/lib/python3.9/site-packages/tensorflow/__init__.py'> | |
| tf.while_loop = <global> <function while_loop_v2 at 0x7f963be03940> | |
| cond = <local> <function _SubnetworkRecCell.get_output.<locals>.cond at 0x7f95ce411c10> | |
| body = <local> <function _SubnetworkRecCell.get_output.<locals>.body at 0x7f95ce411ca0> | |
| loop_vars = <local> (<tf.Tensor 'output/rec/initial_i:0' shape=() dtype=int32>, ([<tf.Tensor 'output/rec/output/init_output_const/constant_with_shape/Mul:0' shape=(?,) dtype=int32>], [[<tf.Tensor 'output/rec/output/zeros:0' shape=(?, 1) dtype=float32>, <tf.Tensor 'output/rec/output/zeros_1:0' shape=(?, 3) dtype=int3... | |
| shape_invariants = <local> (TensorShape([]), ([TensorShape([Dimension(None)])], [[TensorShape([Dimension(None), Dimension(None)]), TensorShape([Dimension(None), Dimension(None)])]]), [TensorShape(None), TensorShape(None)]), _[0]: {len = 0} | |
| back_prop = <not found> | |
| self = <local> <_SubnetworkRecCell '<network via test_concat_new_dim_tag>/output(rec-subnet)'> | |
| self.parent_rec_layer = <local> <RecLayer 'output' out_type=Data{[T|'time'[B&Beam{'output/output'}(3)],B&Beam{'output/output'}(3)], dtype='int32', sparse_dim=DimensionTag{F'classes:sparse-dim'(5)}}> | |
| self.parent_rec_layer.back_prop = <local> False | |
| File "/home/robin/miniconda3/envs/i6_env/lib/python3.9/site-packages/tensorflow/python/util/deprecation.py", line 602, in while_loop_v2 | |
| line: return func(*args, **kwargs) | |
| locals: | |
| func = <local> <function while_loop_v2 at 0x7f963be038b0> | |
| args = <local> () | |
| kwargs = <local> {'cond': <function _SubnetworkRecCell.get_output.<locals>.cond at 0x7f95ce411c10>, 'body': <function _SubnetworkRecCell.get_output.<locals>.body at 0x7f95ce411ca0>, 'loop_vars': (<tf.Tensor 'output/rec/initial_i:0' shape=() dtype=int32>, ([<tf.Tensor 'output/rec/output/init_output_const/constant_... | |
| File "/home/robin/miniconda3/envs/i6_env/lib/python3.9/site-packages/tensorflow/python/ops/control_flow_ops.py", line 2531, in while_loop_v2 | |
| line: return while_loop( | |
| cond=cond, | |
| body=body, | |
| loop_vars=loop_vars, | |
| shape_invariants=shape_invariants, | |
| parallel_iterations=parallel_iterations, | |
| back_prop=back_prop, | |
| swap_memory=swap_memory, | |
| name=name, | |
| maximum_iterations=maximum_iterations, | |
| return_same_structure=True) | |
| locals: | |
| while_loop = <global> <function while_loop at 0x7f963be01790> | |
| cond = <local> <function _SubnetworkRecCell.get_output.<locals>.cond at 0x7f95ce411c10> | |
| body = <local> <function _SubnetworkRecCell.get_output.<locals>.body at 0x7f95ce411ca0> | |
| loop_vars = <local> (<tf.Tensor 'output/rec/initial_i:0' shape=() dtype=int32>, ([<tf.Tensor 'output/rec/output/init_output_const/constant_with_shape/Mul:0' shape=(?,) dtype=int32>], [[<tf.Tensor 'output/rec/output/zeros:0' shape=(?, 1) dtype=float32>, <tf.Tensor 'output/rec/output/zeros_1:0' shape=(?, 3) dtype=int3... | |
| shape_invariants = <local> (TensorShape([]), ([TensorShape([Dimension(None)])], [[TensorShape([Dimension(None), Dimension(None)]), TensorShape([Dimension(None), Dimension(None)])]]), [TensorShape(None), TensorShape(None)]), _[0]: {len = 0} | |
| parallel_iterations = <local> 10 | |
| back_prop = <local> False | |
| swap_memory = <local> False | |
| name = <local> None | |
| maximum_iterations = <local> None | |
| return_same_structure = <not found> | |
| File "/home/robin/miniconda3/envs/i6_env/lib/python3.9/site-packages/tensorflow/python/ops/control_flow_ops.py", line 2815, in while_loop | |
| line: result = loop_context.BuildLoop(cond, body, loop_vars, shape_invariants, | |
| return_same_structure) | |
| locals: | |
| result = <not found> | |
| loop_context = <local> <tensorflow.python.ops.control_flow_ops.WhileContext object at 0x7f95ce3f6100> | |
| loop_context.BuildLoop = <local> <bound method WhileContext.BuildLoop of <tensorflow.python.ops.control_flow_ops.WhileContext object at 0x7f95ce3f6100>> | |
| cond = <local> <function _SubnetworkRecCell.get_output.<locals>.cond at 0x7f95ce411c10> | |
| body = <local> <function _SubnetworkRecCell.get_output.<locals>.body at 0x7f95ce411ca0> | |
| loop_vars = <local> (<tf.Tensor 'output/rec/initial_i:0' shape=() dtype=int32>, ([<tf.Tensor 'output/rec/output/init_output_const/constant_with_shape/Mul:0' shape=(?,) dtype=int32>], [[<tf.Tensor 'output/rec/output/zeros:0' shape=(?, 1) dtype=float32>, <tf.Tensor 'output/rec/output/zeros_1:0' shape=(?, 3) dtype=int3... | |
| shape_invariants = <local> (TensorShape([]), ([TensorShape([Dimension(None)])], [[TensorShape([Dimension(None), Dimension(None)]), TensorShape([Dimension(None), Dimension(None)])]]), [TensorShape(None), TensorShape(None)]), _[0]: {len = 0} | |
| return_same_structure = <local> True | |
| File "/home/robin/miniconda3/envs/i6_env/lib/python3.9/site-packages/tensorflow/python/ops/control_flow_ops.py", line 2297, in WhileContext.BuildLoop | |
| line: original_body_result, exit_vars = self._BuildLoop( | |
| pred, body, original_loop_vars, loop_vars, shape_invariants) | |
| locals: | |
| original_body_result = <not found> | |
| exit_vars = <not found> | |
| self = <local> <tensorflow.python.ops.control_flow_ops.WhileContext object at 0x7f95ce3f6100> | |
| self._BuildLoop = <local> <bound method WhileContext._BuildLoop of <tensorflow.python.ops.control_flow_ops.WhileContext object at 0x7f95ce3f6100>> | |
| pred = <local> <function _SubnetworkRecCell.get_output.<locals>.cond at 0x7f95ce411c10> | |
| body = <local> <function _SubnetworkRecCell.get_output.<locals>.body at 0x7f95ce411ca0> | |
| original_loop_vars = <local> (<tf.Tensor 'output/rec/initial_i:0' shape=() dtype=int32>, ([<tf.Tensor 'output/rec/output/init_output_const/constant_with_shape/Mul:0' shape=(?,) dtype=int32>], [[<tf.Tensor 'output/rec/output/zeros:0' shape=(?, 1) dtype=float32>, <tf.Tensor 'output/rec/output/zeros_1:0' shape=(?, 3) dtype=int3... | |
| loop_vars = <local> [<tf.Tensor 'output/rec/initial_i:0' shape=() dtype=int32>, <tf.Tensor 'output/rec/output/init_output_const/constant_with_shape/Mul:0' shape=(?,) dtype=int32>, <tf.Tensor 'output/rec/output/zeros:0' shape=(?, 1) dtype=float32>, <tf.Tensor 'output/rec/output/zeros_1:0' shape=(?, 3) dtype=int32>, <..., len = 6 | |
| shape_invariants = <local> (TensorShape([]), ([TensorShape([Dimension(None)])], [[TensorShape([Dimension(None), Dimension(None)]), TensorShape([Dimension(None), Dimension(None)])]]), [TensorShape(None), TensorShape(None)]), _[0]: {len = 0} | |
| File "/home/robin/miniconda3/envs/i6_env/lib/python3.9/site-packages/tensorflow/python/ops/control_flow_ops.py", line 2223, in WhileContext._BuildLoop | |
| line: body_result = body(*packed_vars_for_body) | |
| locals: | |
| body_result = <not found> | |
| body = <local> <function _SubnetworkRecCell.get_output.<locals>.body at 0x7f95ce411ca0> | |
| packed_vars_for_body = <local> (<tf.Tensor 'output/rec/while/Identity:0' shape=() dtype=int32>, ([<tf.Tensor 'output/rec/while/Identity_1:0' shape=(?,) dtype=int32>], [[<tf.Tensor 'output/rec/while/Identity_2:0' shape=(?, ?) dtype=float32>, <tf.Tensor 'output/rec/while/Identity_3:0' shape=(?, ?) dtype=int32>]]), [<tf.TensorArr... | |
| File "/mnt/projects/i6/returnn/returnn/tf/layers/rec.py", line 2700, in _SubnetworkRecCell.get_output.<locals>.body | |
| line: self._construct( | |
| prev_outputs=prev_outputs, prev_extra=prev_extra, | |
| i=i, | |
| data=data_, | |
| inputs_moved_out_tas=input_layers_moved_out_tas, | |
| needed_outputs=needed_outputs) | |
| locals: | |
| self = <local> <_SubnetworkRecCell '<network via test_concat_new_dim_tag>/output(rec-subnet)'> | |
| self._construct = <local> <bound method _SubnetworkRecCell._construct of <_SubnetworkRecCell '<network via test_concat_new_dim_tag>/output(rec-subnet)'>> | |
| prev_outputs = <local> {'output': <tf.Tensor 'output/rec/while_loop_body/prev_outputs/identity_output:0' shape=(?,) dtype=int32>} | |
| prev_extra = <local> {'output': {'choice_scores': <tf.Tensor 'output/rec/while_loop_body/prev_extra/identity_output_choice_scores:0' shape=(?, ?) dtype=float32>, 'choice_src_beams': <tf.Tensor 'output/rec/while_loop_body/prev_extra/identity_output_choice_src_beams:0' shape=(?, ?) dtype=int32>}} | |
| i = <local> <tf.Tensor 'output/rec/while/Identity:0' shape=() dtype=int32> | |
| data = <not found> | |
| data_ = <local> {'source': <tf.Tensor 'output/rec/while_loop_body/source_ta_read:0' shape=(?, 5) dtype=float32>} | |
| inputs_moved_out_tas = <not found> | |
| input_layers_moved_out_tas = <local> {} | |
| needed_outputs = <local> {'output'}, len = 1 | |
| File "/mnt/projects/i6/returnn/returnn/tf/layers/rec.py", line 1813, in _SubnetworkRecCell._construct | |
| line: layer = get_layer(layer_name) | |
| locals: | |
| layer = <not found> | |
| get_layer = <local> <function _SubnetworkRecCell._construct.<locals>.get_layer at 0x7f95ce411b80> | |
| layer_name = <local> 'output', len = 6 | |
| File "/mnt/projects/i6/returnn/returnn/tf/layers/rec.py", line 1781, in _SubnetworkRecCell._construct.<locals>.get_layer | |
| line: layer = self.net.construct_layer(self.net_dict, name=name, get_layer=get_layer) | |
| locals: | |
| layer = <not found> | |
| self = <local> <_SubnetworkRecCell '<network via test_concat_new_dim_tag>/output(rec-subnet)'> | |
| self.net = <local> <TFNetwork '<network via test_concat_new_dim_tag>/output(rec-subnet)' parent_layer=<RecLayer 'output' out_type=Data{[T|'time'[B&Beam{'output/output'}(3)],B&Beam{'output/output'}(3)], dtype='int32', sparse_dim=DimensionTag{F'classes:sparse-dim'(5)}}> train=False search> | |
| self.net.construct_layer = <local> <bound method TFNetwork.construct_layer of <TFNetwork '<network via test_concat_new_dim_tag>/output(rec-subnet)' parent_layer=<RecLayer 'output' out_type=Data{[T|'time'[B&Beam{'output/output'}(3)],B&Beam{'output/output'}(3)], dtype='int32', sparse_dim=DimensionTag{F'classes:sparse-dim'(5)}}> trai... | |
| self.net_dict = <local> {'prev_out0': {'class': 'reinterpret_data', 'from': 'prev:output', 'set_sparse': False}, 'prev_out1': {'class': 'cast', 'from': 'prev_out0', 'dtype': 'float32'}, 'prev_out': {'class': 'expand_dims', 'from': 'prev_out1', 'axis': 'f'}, 'data_concat': {'class': 'copy', 'from': ['base:data_new', 'pre..., len = 7 | |
| name = <local> 'output', len = 6 | |
| get_layer = <local> <function _SubnetworkRecCell._construct.<locals>.get_layer at 0x7f95ce411b80> | |
| File "/mnt/projects/i6/returnn/returnn/tf/network.py", line 928, in TFNetwork.construct_layer | |
| line: layer_class.transform_config_dict(layer_desc, network=net, get_layer=get_layer) | |
| locals: | |
| layer_class = <local> <class 'returnn.tf.layers.rec.ChoiceLayer'> | |
| layer_class.transform_config_dict = <local> <bound method ChoiceLayer.transform_config_dict of <class 'returnn.tf.layers.rec.ChoiceLayer'>> | |
| layer_desc = <local> {'beam_size': 3, 'target': ['classes'], 'input_type': 'prob', 'initial_output': 0, '_network': <TFNetwork '<network via test_concat_new_dim_tag>/output(rec-subnet)' parent_layer=<RecLayer 'output' out_type=Data{[T|'time'[B&Beam{'output/output'}(3)],B&Beam{'output/output'}(3)], dtype='int32', spar..., len = 6 | |
| network = <not found> | |
| net = <local> <TFNetwork '<network via test_concat_new_dim_tag>/output(rec-subnet)' parent_layer=<RecLayer 'output' out_type=Data{[T|'time'[B&Beam{'output/output'}(3)],B&Beam{'output/output'}(3)], dtype='int32', sparse_dim=DimensionTag{F'classes:sparse-dim'(5)}}> train=False search> | |
| get_layer = <local> <function _SubnetworkRecCell._construct.<locals>.get_layer at 0x7f95ce411b80> | |
| File "/mnt/projects/i6/returnn/returnn/tf/layers/rec.py", line 5579, in ChoiceLayer.transform_config_dict | |
| line: super(ChoiceLayer, cls).transform_config_dict(d, network=network, get_layer=get_layer) | |
| locals: | |
| super = <builtin> <class 'super'> | |
| ChoiceLayer = <global> <class 'returnn.tf.layers.rec.ChoiceLayer'> | |
| cls = <local> <class 'returnn.tf.layers.rec.ChoiceLayer'> | |
| transform_config_dict = <not found> | |
| d = <local> {'beam_size': 3, 'target': ['classes'], 'input_type': 'prob', 'initial_output': 0, '_network': <TFNetwork '<network via test_concat_new_dim_tag>/output(rec-subnet)' parent_layer=<RecLayer 'output' out_type=Data{[T|'time'[B&Beam{'output/output'}(3)],B&Beam{'output/output'}(3)], dtype='int32', spar..., len = 6 | |
| network = <local> <TFNetwork '<network via test_concat_new_dim_tag>/output(rec-subnet)' parent_layer=<RecLayer 'output' out_type=Data{[T|'time'[B&Beam{'output/output'}(3)],B&Beam{'output/output'}(3)], dtype='int32', sparse_dim=DimensionTag{F'classes:sparse-dim'(5)}}> train=False search> | |
| get_layer = <local> <function _SubnetworkRecCell._construct.<locals>.get_layer at 0x7f95ce411b80> | |
| File "/mnt/projects/i6/returnn/returnn/tf/layers/rec.py", line 4991, in BaseChoiceLayer.transform_config_dict | |
| line: super(BaseChoiceLayer, cls).transform_config_dict(d, network=network, get_layer=get_layer) | |
| locals: | |
| super = <builtin> <class 'super'> | |
| BaseChoiceLayer = <global> <class 'returnn.tf.layers.rec.BaseChoiceLayer'> | |
| cls = <local> <class 'returnn.tf.layers.rec.ChoiceLayer'> | |
| transform_config_dict = <not found> | |
| d = <local> {'beam_size': 3, 'target': ['classes'], 'input_type': 'prob', 'initial_output': 0, '_network': <TFNetwork '<network via test_concat_new_dim_tag>/output(rec-subnet)' parent_layer=<RecLayer 'output' out_type=Data{[T|'time'[B&Beam{'output/output'}(3)],B&Beam{'output/output'}(3)], dtype='int32', spar..., len = 6 | |
| network = <local> <TFNetwork '<network via test_concat_new_dim_tag>/output(rec-subnet)' parent_layer=<RecLayer 'output' out_type=Data{[T|'time'[B&Beam{'output/output'}(3)],B&Beam{'output/output'}(3)], dtype='int32', sparse_dim=DimensionTag{F'classes:sparse-dim'(5)}}> train=False search> | |
| get_layer = <local> <function _SubnetworkRecCell._construct.<locals>.get_layer at 0x7f95ce411b80> | |
| File "/mnt/projects/i6/returnn/returnn/tf/layers/base.py", line 562, in LayerBase.transform_config_dict | |
| line: d["sources"] = [ | |
| get_layer(src_name) | |
| for src_name in src_names | |
| if not src_name == "none"] | |
| locals: | |
| d = <local> {'beam_size': 3, 'target': ['classes'], 'input_type': 'prob', 'initial_output': 0, '_network': <TFNetwork '<network via test_concat_new_dim_tag>/output(rec-subnet)' parent_layer=<RecLayer 'output' out_type=Data{[T|'time'[B&Beam{'output/output'}(3)],B&Beam{'output/output'}(3)], dtype='int32', spar..., len = 6 | |
| get_layer = <local> <function _SubnetworkRecCell._construct.<locals>.get_layer at 0x7f95ce411b80> | |
| src_name = <not found> | |
| src_names = <local> ['output_prob'], _[0]: {len = 11} | |
| File "/mnt/projects/i6/returnn/returnn/tf/layers/base.py", line 563, in <listcomp> | |
| line: get_layer(src_name) | |
| locals: | |
| get_layer = <local> <function _SubnetworkRecCell._construct.<locals>.get_layer at 0x7f95ce411b80> | |
| src_name = <local> 'output_prob', len = 11 | |
| File "/mnt/projects/i6/returnn/returnn/tf/layers/rec.py", line 1781, in _SubnetworkRecCell._construct.<locals>.get_layer | |
| line: layer = self.net.construct_layer(self.net_dict, name=name, get_layer=get_layer) | |
| locals: | |
| layer = <not found> | |
| self = <local> <_SubnetworkRecCell '<network via test_concat_new_dim_tag>/output(rec-subnet)'> | |
| self.net = <local> <TFNetwork '<network via test_concat_new_dim_tag>/output(rec-subnet)' parent_layer=<RecLayer 'output' out_type=Data{[T|'time'[B&Beam{'output/output'}(3)],B&Beam{'output/output'}(3)], dtype='int32', sparse_dim=DimensionTag{F'classes:sparse-dim'(5)}}> train=False search> | |
| self.net.construct_layer = <local> <bound method TFNetwork.construct_layer of <TFNetwork '<network via test_concat_new_dim_tag>/output(rec-subnet)' parent_layer=<RecLayer 'output' out_type=Data{[T|'time'[B&Beam{'output/output'}(3)],B&Beam{'output/output'}(3)], dtype='int32', sparse_dim=DimensionTag{F'classes:sparse-dim'(5)}}> trai... | |
| self.net_dict = <local> {'prev_out0': {'class': 'reinterpret_data', 'from': 'prev:output', 'set_sparse': False}, 'prev_out1': {'class': 'cast', 'from': 'prev_out0', 'dtype': 'float32'}, 'prev_out': {'class': 'expand_dims', 'from': 'prev_out1', 'axis': 'f'}, 'data_concat': {'class': 'copy', 'from': ['base:data_new', 'pre..., len = 7 | |
| name = <local> 'output_prob', len = 11 | |
| get_layer = <local> <function _SubnetworkRecCell._construct.<locals>.get_layer at 0x7f95ce411b80> | |
| File "/mnt/projects/i6/returnn/returnn/tf/network.py", line 928, in TFNetwork.construct_layer | |
| line: layer_class.transform_config_dict(layer_desc, network=net, get_layer=get_layer) | |
| locals: | |
| layer_class = <local> <class 'returnn.tf.layers.basic.SoftmaxLayer'> | |
| layer_class.transform_config_dict = <local> <bound method LayerBase.transform_config_dict of <class 'returnn.tf.layers.basic.SoftmaxLayer'>> | |
| layer_desc = <local> {'target': 'classes', 'loss': 'ce', '_network': <TFNetwork '<network via test_concat_new_dim_tag>/output(rec-subnet)' parent_layer=<RecLayer 'output' out_type=Data{[T|'time'[B&Beam{'output/output'}(3)],B&Beam{'output/output'}(3)], dtype='int32', sparse_dim=DimensionTag{F'classes:sparse-dim'(5)}}>... | |
| network = <not found> | |
| net = <local> <TFNetwork '<network via test_concat_new_dim_tag>/output(rec-subnet)' parent_layer=<RecLayer 'output' out_type=Data{[T|'time'[B&Beam{'output/output'}(3)],B&Beam{'output/output'}(3)], dtype='int32', sparse_dim=DimensionTag{F'classes:sparse-dim'(5)}}> train=False search> | |
| get_layer = <local> <function _SubnetworkRecCell._construct.<locals>.get_layer at 0x7f95ce411b80> | |
| File "/mnt/projects/i6/returnn/returnn/tf/layers/base.py", line 562, in LayerBase.transform_config_dict | |
| line: d["sources"] = [ | |
| get_layer(src_name) | |
| for src_name in src_names | |
| if not src_name == "none"] | |
| locals: | |
| d = <local> {'target': 'classes', 'loss': 'ce', '_network': <TFNetwork '<network via test_concat_new_dim_tag>/output(rec-subnet)' parent_layer=<RecLayer 'output' out_type=Data{[T|'time'[B&Beam{'output/output'}(3)],B&Beam{'output/output'}(3)], dtype='int32', sparse_dim=DimensionTag{F'classes:sparse-dim'(5)}}>... | |
| get_layer = <local> <function _SubnetworkRecCell._construct.<locals>.get_layer at 0x7f95ce411b80> | |
| src_name = <not found> | |
| src_names = <local> ['data_red'], _[0]: {len = 8} | |
| File "/mnt/projects/i6/returnn/returnn/tf/layers/base.py", line 563, in <listcomp> | |
| line: get_layer(src_name) | |
| locals: | |
| get_layer = <local> <function _SubnetworkRecCell._construct.<locals>.get_layer at 0x7f95ce411b80> | |
| src_name = <local> 'data_red', len = 8 | |
| File "/mnt/projects/i6/returnn/returnn/tf/layers/rec.py", line 1781, in _SubnetworkRecCell._construct.<locals>.get_layer | |
| line: layer = self.net.construct_layer(self.net_dict, name=name, get_layer=get_layer) | |
| locals: | |
| layer = <not found> | |
| self = <local> <_SubnetworkRecCell '<network via test_concat_new_dim_tag>/output(rec-subnet)'> | |
| self.net = <local> <TFNetwork '<network via test_concat_new_dim_tag>/output(rec-subnet)' parent_layer=<RecLayer 'output' out_type=Data{[T|'time'[B&Beam{'output/output'}(3)],B&Beam{'output/output'}(3)], dtype='int32', sparse_dim=DimensionTag{F'classes:sparse-dim'(5)}}> train=False search> | |
| self.net.construct_layer = <local> <bound method TFNetwork.construct_layer of <TFNetwork '<network via test_concat_new_dim_tag>/output(rec-subnet)' parent_layer=<RecLayer 'output' out_type=Data{[T|'time'[B&Beam{'output/output'}(3)],B&Beam{'output/output'}(3)], dtype='int32', sparse_dim=DimensionTag{F'classes:sparse-dim'(5)}}> trai... | |
| self.net_dict = <local> {'prev_out0': {'class': 'reinterpret_data', 'from': 'prev:output', 'set_sparse': False}, 'prev_out1': {'class': 'cast', 'from': 'prev_out0', 'dtype': 'float32'}, 'prev_out': {'class': 'expand_dims', 'from': 'prev_out1', 'axis': 'f'}, 'data_concat': {'class': 'copy', 'from': ['base:data_new', 'pre..., len = 7 | |
| name = <local> 'data_red', len = 8 | |
| get_layer = <local> <function _SubnetworkRecCell._construct.<locals>.get_layer at 0x7f95ce411b80> | |
| File "/mnt/projects/i6/returnn/returnn/tf/network.py", line 928, in TFNetwork.construct_layer | |
| line: layer_class.transform_config_dict(layer_desc, network=net, get_layer=get_layer) | |
| locals: | |
| layer_class = <local> <class 'returnn.tf.layers.basic.ReduceLayer'> | |
| layer_class.transform_config_dict = <local> <bound method LayerBase.transform_config_dict of <class 'returnn.tf.layers.basic.ReduceLayer'>> | |
| layer_desc = <local> {'axis': 'stag:new-time', 'mode': 'max', '_network': <TFNetwork '<network via test_concat_new_dim_tag>/output(rec-subnet)' parent_layer=<RecLayer 'output' out_type=Data{[T|'time'[B&Beam{'output/output'}(3)],B&Beam{'output/output'}(3)], dtype='int32', sparse_dim=DimensionTag{F'classes:sparse-dim'(... | |
| network = <not found> | |
| net = <local> <TFNetwork '<network via test_concat_new_dim_tag>/output(rec-subnet)' parent_layer=<RecLayer 'output' out_type=Data{[T|'time'[B&Beam{'output/output'}(3)],B&Beam{'output/output'}(3)], dtype='int32', sparse_dim=DimensionTag{F'classes:sparse-dim'(5)}}> train=False search> | |
| get_layer = <local> <function _SubnetworkRecCell._construct.<locals>.get_layer at 0x7f95ce411b80> | |
| File "/mnt/projects/i6/returnn/returnn/tf/layers/base.py", line 562, in LayerBase.transform_config_dict | |
| line: d["sources"] = [ | |
| get_layer(src_name) | |
| for src_name in src_names | |
| if not src_name == "none"] | |
| locals: | |
| d = <local> {'axis': 'stag:new-time', 'mode': 'max', '_network': <TFNetwork '<network via test_concat_new_dim_tag>/output(rec-subnet)' parent_layer=<RecLayer 'output' out_type=Data{[T|'time'[B&Beam{'output/output'}(3)],B&Beam{'output/output'}(3)], dtype='int32', sparse_dim=DimensionTag{F'classes:sparse-dim'(... | |
| get_layer = <local> <function _SubnetworkRecCell._construct.<locals>.get_layer at 0x7f95ce411b80> | |
| src_name = <not found> | |
| src_names = <local> ['data_concat'], _[0]: {len = 11} | |
| File "/mnt/projects/i6/returnn/returnn/tf/layers/base.py", line 563, in <listcomp> | |
| line: get_layer(src_name) | |
| locals: | |
| get_layer = <local> <function _SubnetworkRecCell._construct.<locals>.get_layer at 0x7f95ce411b80> | |
| src_name = <local> 'data_concat', len = 11 | |
| File "/mnt/projects/i6/returnn/returnn/tf/layers/rec.py", line 1781, in _SubnetworkRecCell._construct.<locals>.get_layer | |
| line: layer = self.net.construct_layer(self.net_dict, name=name, get_layer=get_layer) | |
| locals: | |
| layer = <not found> | |
| self = <local> <_SubnetworkRecCell '<network via test_concat_new_dim_tag>/output(rec-subnet)'> | |
| self.net = <local> <TFNetwork '<network via test_concat_new_dim_tag>/output(rec-subnet)' parent_layer=<RecLayer 'output' out_type=Data{[T|'time'[B&Beam{'output/output'}(3)],B&Beam{'output/output'}(3)], dtype='int32', sparse_dim=DimensionTag{F'classes:sparse-dim'(5)}}> train=False search> | |
| self.net.construct_layer = <local> <bound method TFNetwork.construct_layer of <TFNetwork '<network via test_concat_new_dim_tag>/output(rec-subnet)' parent_layer=<RecLayer 'output' out_type=Data{[T|'time'[B&Beam{'output/output'}(3)],B&Beam{'output/output'}(3)], dtype='int32', sparse_dim=DimensionTag{F'classes:sparse-dim'(5)}}> trai... | |
| self.net_dict = <local> {'prev_out0': {'class': 'reinterpret_data', 'from': 'prev:output', 'set_sparse': False}, 'prev_out1': {'class': 'cast', 'from': 'prev_out0', 'dtype': 'float32'}, 'prev_out': {'class': 'expand_dims', 'from': 'prev_out1', 'axis': 'f'}, 'data_concat': {'class': 'copy', 'from': ['base:data_new', 'pre..., len = 7 | |
| name = <local> 'data_concat', len = 11 | |
| get_layer = <local> <function _SubnetworkRecCell._construct.<locals>.get_layer at 0x7f95ce411b80> | |
| File "/mnt/projects/i6/returnn/returnn/tf/network.py", line 935, in TFNetwork.construct_layer | |
| line: return add_layer(name=name_with_prefix, layer_class=layer_class, **layer_desc) | |
| locals: | |
| add_layer = <local> <bound method TFNetwork.add_layer of <TFNetwork '<network via test_concat_new_dim_tag>/output(rec-subnet)' parent_layer=<RecLayer 'output' out_type=Data{[T|'time'[B&Beam{'output/output'}(3)],B&Beam{'output/output'}(3)], dtype='int32', sparse_dim=DimensionTag{F'classes:sparse-dim'(5)}}> train=Fals... | |
| name = <local> 'data_concat', len = 11 | |
| name_with_prefix = <local> 'data_concat', len = 11 | |
| layer_class = <local> <class 'returnn.tf.layers.basic.CopyLayer'> | |
| layer_desc = <local> {'_network': <TFNetwork '<network via test_concat_new_dim_tag>/output(rec-subnet)' parent_layer=<RecLayer 'output' out_type=Data{[T|'time'[B&Beam{'output/output'}(3)],B&Beam{'output/output'}(3)], dtype='int32', sparse_dim=DimensionTag{F'classes:sparse-dim'(5)}}> train=False search>, '_name': 'dat... | |
| File "/mnt/projects/i6/returnn/returnn/tf/network.py", line 1082, in TFNetwork.add_layer | |
| line: layer = self._create_layer(name=name, layer_class=layer_class, **layer_desc) | |
| locals: | |
| layer = <not found> | |
| self = <local> <TFNetwork '<network via test_concat_new_dim_tag>/output(rec-subnet)' parent_layer=<RecLayer 'output' out_type=Data{[T|'time'[B&Beam{'output/output'}(3)],B&Beam{'output/output'}(3)], dtype='int32', sparse_dim=DimensionTag{F'classes:sparse-dim'(5)}}> train=False search> | |
| self._create_layer = <local> <bound method TFNetwork._create_layer of <TFNetwork '<network via test_concat_new_dim_tag>/output(rec-subnet)' parent_layer=<RecLayer 'output' out_type=Data{[T|'time'[B&Beam{'output/output'}(3)],B&Beam{'output/output'}(3)], dtype='int32', sparse_dim=DimensionTag{F'classes:sparse-dim'(5)}}> train=... | |
| name = <local> 'data_concat', len = 11 | |
| layer_class = <local> <class 'returnn.tf.layers.basic.CopyLayer'> | |
| layer_desc = <local> {'_network': <TFNetwork '<network via test_concat_new_dim_tag>/output(rec-subnet)' parent_layer=<RecLayer 'output' out_type=Data{[T|'time'[B&Beam{'output/output'}(3)],B&Beam{'output/output'}(3)], dtype='int32', sparse_dim=DimensionTag{F'classes:sparse-dim'(5)}}> train=False search>, '_name': 'dat... | |
| File "/mnt/projects/i6/returnn/returnn/tf/network.py", line 1002, in TFNetwork._create_layer | |
| line: layer = layer_class(**layer_desc) | |
| locals: | |
| layer = <not found> | |
| layer_class = <local> <class 'returnn.tf.layers.basic.CopyLayer'> | |
| layer_desc = <local> {'_network': <TFNetwork '<network via test_concat_new_dim_tag>/output(rec-subnet)' parent_layer=<RecLayer 'output' out_type=Data{[T|'time'[B&Beam{'output/output'}(3)],B&Beam{'output/output'}(3)], dtype='int32', sparse_dim=DimensionTag{F'classes:sparse-dim'(5)}}> train=False search>, '_name': 'dat..., len = 7 | |
| File "/mnt/projects/i6/returnn/returnn/tf/layers/basic.py", line 293, in CopyLayer.__init__ | |
| line: super(CopyLayer, self).__init__(in_dim=in_dim, out_dim=out_dim, **kwargs) | |
| locals: | |
| super = <builtin> <class 'super'> | |
| CopyLayer = <global> <class 'returnn.tf.layers.basic.CopyLayer'> | |
| self = <local> <CopyLayer output/'data_concat' out_type=Data{[B&Beam{'output/prev:output'}(3),T|'new-time'[B&Beam{'output/prev:output'}(3)],F|F'data_concat_output_feature'(6)], ctx=loop('time'[B])}> | |
| __init__ = <not found> | |
| in_dim = <local> None | |
| out_dim = <local> None | |
| kwargs = <local> {'_network': <TFNetwork '<network via test_concat_new_dim_tag>/output(rec-subnet)' parent_layer=<RecLayer 'output' out_type=Data{[T|'time'[B&Beam{'output/output'}(3)],B&Beam{'output/output'}(3)], dtype='int32', sparse_dim=DimensionTag{F'classes:sparse-dim'(5)}}> train=False search>, '_name': 'dat..., len = 7 | |
| File "/mnt/projects/i6/returnn/returnn/tf/layers/basic.py", line 265, in _ConcatInputLayer.__init__ | |
| line: self.input_data = concat_sources_with_opt_dropout( | |
| self.sources, out_dim=in_dim, | |
| dropout=dropout, dropout_noise_shape=dropout_noise_shape, dropout_on_forward=dropout_on_forward, | |
| allow_broadcast_all_sources=True if out_shape else NotSpecified) | |
| locals: | |
| self = <local> <CopyLayer output/'data_concat' out_type=Data{[B&Beam{'output/prev:output'}(3),T|'new-time'[B&Beam{'output/prev:output'}(3)],F|F'data_concat_output_feature'(6)], ctx=loop('time'[B])}> | |
| self.input_data = <local> None | |
| concat_sources_with_opt_dropout = <global> <function concat_sources_with_opt_dropout at 0x7f95cde40550> | |
| self.sources = <local> [<SelectSearchSourcesLayer 'data_new' <SearchChoices owner='prev:output' beam_size=3 beam_scores=shaped:(None,None)> out_type=Data{[B&Beam{'output/prev:output'}(3),T|'new-time'[B&Beam{'output/prev:output'}(3)],F|F'feature:data'(5)]}>, <ExpandDimsLayer output/'prev_out' out_type=Data{[B&Beam{'outp... | |
| out_dim = <not found> | |
| in_dim = <local> None | |
| dropout = <local> 0 | |
| dropout_noise_shape = <local> None | |
| dropout_on_forward = <local> False | |
| allow_broadcast_all_sources = <not found> | |
| out_shape = <local> None | |
| NotSpecified = <global> <class 'returnn.util.basic.NotSpecified'> | |
| File "/mnt/projects/i6/returnn/returnn/tf/layers/basic.py", line 201, in concat_sources_with_opt_dropout | |
| line: data = concat_sources(src_layers, out_dim=out_dim, allow_broadcast_all_sources=allow_broadcast_all_sources) | |
| locals: | |
| data = <not found> | |
| concat_sources = <global> <function concat_sources at 0x7f95cde40430> | |
| src_layers = <local> [<SelectSearchSourcesLayer 'data_new' <SearchChoices owner='prev:output' beam_size=3 beam_scores=shaped:(None,None)> out_type=Data{[B&Beam{'output/prev:output'}(3),T|'new-time'[B&Beam{'output/prev:output'}(3)],F|F'feature:data'(5)]}>, <ExpandDimsLayer output/'prev_out' out_type=Data{[B&Beam{'outp... | |
| out_dim = <local> None | |
| allow_broadcast_all_sources = <local> <class 'returnn.util.basic.NotSpecified'> | |
| File "/mnt/projects/i6/returnn/returnn/tf/layers/basic.py", line 126, in concat_sources | |
| line: data.placeholder = tf.concat( | |
| axis=data.feature_dim_axis, | |
| values=[layer_data.placeholder for layer_data in layers_data]) | |
| locals: | |
| data = <local> Data{'concat_data_new_prev_out', [B&Beam{'output/prev:output'}(3),T|'new-time'[B],F|F'concat_data_new_prev_out_feature'(6)]} | |
| data.placeholder = <local> <tf.Tensor 'concat_data_new_prev_out/concat_sources/concat:0' shape=(?, 1, 6) dtype=float32> | |
| tf = <global> <module 'tensorflow' from '/home/robin/miniconda3/envs/i6_env/lib/python3.9/site-packages/tensorflow/__init__.py'> | |
| tf.concat = <global> <function concat at 0x7f964c7275e0> | |
| axis = <not found> | |
| data.feature_dim_axis = <local> 2 | |
| values = <not found> | |
| layer_data = <not found> | |
| layer_data.placeholder = <not found> | |
| layers_data = <local> [Data{'data_new_output', [B&Beam{'output/prev:output'}(3),T|'new-time'[B&Beam{'output/prev:output'}(3)],F|F'feature:data'(5)]}, Data{'prev_out_output', [B&Beam{'output/prev:output'}(3),T|'prev_out_expand_dims'(1),F|F'concat_data_new_prev_out_feature_dummy_dim1'(1)], ctx=loop('time'[B])}] | |
| File "/mnt/projects/i6/returnn/returnn/tf/util/data.py", line 3256, in Data.placeholder | |
| line: self.sanity_check(assume_complete=False) | |
| locals: | |
| self = <local> Data{'concat_data_new_prev_out', [B&Beam{'output/prev:output'}(3),T|'new-time'[B],F|F'concat_data_new_prev_out_feature'(6)]} | |
| self.sanity_check = <local> <bound method Data.sanity_check of Data{'concat_data_new_prev_out', [B&Beam{'output/prev:output'}(3),T|'new-time'[B],F|F'concat_data_new_prev_out_feature'(6)]}> | |
| assume_complete = <not found> | |
| File "/mnt/projects/i6/returnn/returnn/tf/util/data.py", line 1820, in Data.sanity_check | |
| line: assert tag.batch == self.batch or self.batch.is_broadcast() | |
| locals: | |
| tag = <local> DimensionTag{'new-time'[B]} | |
| tag.batch = <local> BatchInfo{B} | |
| self = <local> Data{'concat_data_new_prev_out', [B&Beam{'output/prev:output'}(3),T|'new-time'[B],F|F'concat_data_new_prev_out_feature'(6)]} | |
| self.batch = <local> BatchInfo{B, Beam{'output/prev:output'}(3)} | |
| self.batch.is_broadcast = <local> <bound method BatchInfo.is_broadcast of BatchInfo{B, Beam{'output/prev:output'}(3)}> | |
| AssertionError |
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
| network = { | |
| "data_new": {"class": "reinterpret_data", "from": "data", | |
| "set_dim_tags": {"t": new_time_tag}}, | |
| "output": {"class": "rec", "from": "data", "unit": { | |
| "prev_out0": { | |
| "class": "reinterpret_data", "from": "prev:output", "set_sparse": False}, | |
| "prev_out1": {"class": "cast", "from": "prev_out0", "dtype": "float32"}, | |
| "prev_out": {"class": "expand_dims", "from": "prev_out1", "axis": "f"}, | |
| "data_concat": { | |
| "class": "copy", "from": ["base:data_new", "prev_out"] | |
| }, | |
| "data_red": {"class": "reduce", "from": "data_concat", "axis": "stag:new-time", "mode": "max"}, | |
| "output_prob": {"class": "softmax", "from": "data_red", "target": "classes", "loss": "ce"}, | |
| "output": { | |
| "class": "choice", "from": "output_prob", "beam_size": 3, "target": "classes", | |
| "input_type": "prob", "initial_output": 0} | |
| }} | |
| } |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment