/layer-norm.mlir Secret
Created
March 28, 2022 20:37
Star
You must be signed in to star a gist
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
#map0 = affine_map<(d0, d1, d2, d3, d4) -> (d0, d1, d2, d3, d4)> | |
#map1 = affine_map<(d0, d1, d2, d3, d4) -> (d0, d1)> | |
#map2 = affine_map<(d0, d1) -> (d0, d1)> | |
#map3 = affine_map<(d0, d1, d2, d3, d4) -> (d2, d3, d4)> | |
module attributes {torch.debug_module_name = "LayerNormModule"} { | |
func @forward(%arg0: tensor<2x5x2x2x3xf32>) -> tensor<2x5x2x2x3xf32> { | |
%cst = arith.constant 0.000000e+00 : f32 | |
%c3_i64 = arith.constant 3 : i64 | |
%c2_i64 = arith.constant 2 : i64 | |
%cst_0 = arith.constant 1.000000e-05 : f64 | |
%cst_1 = arith.constant dense<[[[3.000000e+00, 2.000000e+00, 4.000000e+00], [2.000000e+00, 3.000000e+00, 3.000000e+00]], [[3.000000e+00, 2.000000e+00, 4.000000e+00], [2.000000e+00, 3.000000e+00, 3.000000e+00]]]> : tensor<2x2x3xf32> | |
%cst_2 = arith.constant dense<[[[5.000000e-01, 4.000000e-01, 3.000000e-01], [2.000000e-01, 4.000000e-01, 3.000000e-01]], [[5.000000e-01, 4.000000e-01, 3.000000e-01], [2.000000e-01, 4.000000e-01, 3.000000e-01]]]> : tensor<2x2x3xf32> | |
%0 = arith.cmpi eq, %c2_i64, %c2_i64 : i64 | |
cf.assert %0, "mismatching contracting dimension" | |
cf.assert %0, "mismatching contracting dimension" | |
cf.assert %0, "mismatching contracting dimension" | |
cf.assert %0, "mismatching contracting dimension" | |
cf.assert %0, "mismatching contracting dimension" | |
cf.assert %0, "mismatching contracting dimension" | |
%1 = arith.cmpi eq, %c3_i64, %c3_i64 : i64 | |
cf.assert %1, "mismatching contracting dimension" | |
cf.assert %1, "mismatching contracting dimension" | |
cf.assert %1, "mismatching contracting dimension" | |
%2 = arith.muli %c2_i64, %c2_i64 : i64 | |
%3 = arith.muli %2, %c3_i64 : i64 | |
%4 = arith.sitofp %3 : i64 to f32 | |
%5 = linalg.init_tensor [2, 5] : tensor<2x5xf32> | |
%6 = linalg.fill ins(%cst : f32) outs(%5 : tensor<2x5xf32>) -> tensor<2x5xf32> | |
%7 = linalg.generic {indexing_maps = [#map0, #map1], iterator_types = ["parallel", "parallel", "reduction", "reduction", "reduction"]} ins(%arg0 : tensor<2x5x2x2x3xf32>) outs(%6 : tensor<2x5xf32>) { | |
^bb0(%arg1: f32, %arg2: f32): | |
%15 = arith.addf %arg2, %arg1 : f32 | |
linalg.yield %15 : f32 | |
} -> tensor<2x5xf32> | |
%8 = linalg.generic {indexing_maps = [#map2, #map2], iterator_types = ["parallel", "parallel"]} ins(%7 : tensor<2x5xf32>) outs(%5 : tensor<2x5xf32>) { | |
^bb0(%arg1: f32, %arg2: f32): | |
%15 = arith.divf %arg1, %4 : f32 | |
linalg.yield %15 : f32 | |
} -> tensor<2x5xf32> | |
%9 = linalg.fill ins(%cst : f32) outs(%5 : tensor<2x5xf32>) -> tensor<2x5xf32> | |
%10 = linalg.generic {indexing_maps = [#map0, #map1, #map1], iterator_types = ["parallel", "parallel", "reduction", "reduction", "reduction"]} ins(%arg0, %8 : tensor<2x5x2x2x3xf32>, tensor<2x5xf32>) outs(%9 : tensor<2x5xf32>) { | |
^bb0(%arg1: f32, %arg2: f32, %arg3: f32): | |
%15 = arith.subf %arg1, %arg2 : f32 | |
%16 = arith.mulf %15, %15 : f32 | |
%17 = arith.addf %arg3, %16 : f32 | |
linalg.yield %17 : f32 | |
} -> tensor<2x5xf32> | |
%11 = linalg.generic {indexing_maps = [#map2, #map2], iterator_types = ["parallel", "parallel"]} ins(%10 : tensor<2x5xf32>) outs(%5 : tensor<2x5xf32>) { | |
^bb0(%arg1: f32, %arg2: f32): | |
%15 = arith.divf %arg1, %4 : f32 | |
linalg.yield %15 : f32 | |
} -> tensor<2x5xf32> | |
%12 = linalg.generic {indexing_maps = [#map2, #map2], iterator_types = ["parallel", "parallel"]} ins(%11 : tensor<2x5xf32>) outs(%5 : tensor<2x5xf32>) { | |
^bb0(%arg1: f32, %arg2: f32): | |
%15 = arith.truncf %cst_0 : f64 to f32 | |
%16 = arith.addf %arg1, %15 : f32 | |
%17 = math.rsqrt %16 : f32 | |
linalg.yield %17 : f32 | |
} -> tensor<2x5xf32> | |
%13 = linalg.init_tensor [2, 5, 2, 2, 3] : tensor<2x5x2x2x3xf32> | |
%14 = linalg.generic {indexing_maps = [#map0, #map1, #map1, #map3, #map3, #map0], iterator_types = ["parallel", "parallel", "parallel", "parallel", "parallel"]} ins(%arg0, %8, %12, %cst_1, %cst_2 : tensor<2x5x2x2x3xf32>, tensor<2x5xf32>, tensor<2x5xf32>, tensor<2x2x3xf32>, tensor<2x2x3xf32>) outs(%13 : tensor<2x5x2x2x3xf32>) { | |
^bb0(%arg1: f32, %arg2: f32, %arg3: f32, %arg4: f32, %arg5: f32, %arg6: f32): | |
%15 = arith.subf %arg1, %arg2 : f32 | |
%16 = arith.mulf %15, %arg3 : f32 | |
%17 = arith.mulf %16, %arg4 : f32 | |
%18 = arith.addf %17, %arg5 : f32 | |
linalg.yield %18 : f32 | |
} -> tensor<2x5x2x2x3xf32> | |
return %14 : tensor<2x5x2x2x3xf32> | |
} | |
} |
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
#map0 = affine_map<(d0, d1, d2) -> (d0, d1, d2)> | |
#map1 = affine_map<(d0, d1, d2) -> (d0, d1, 0)> | |
module attributes {torch.debug_module_name = "SoftmaxIntModule"} { | |
func @forward(%arg0: tensor<?x?x?xf32>) -> tensor<?x?x?xf32> { | |
%c1 = arith.constant 1 : index | |
%c0 = arith.constant 0 : index | |
%c2 = arith.constant 2 : index | |
%cst = arith.constant 0.000000e+00 : f32 | |
%cst_0 = arith.constant 1.000000e+00 : f64 | |
%cst_1 = arith.constant -3.40282347E+38 : f32 | |
%c0_i64 = arith.constant 0 : i64 | |
%0 = tensor.dim %arg0, %c0 : tensor<?x?x?xf32> | |
%1 = tensor.dim %arg0, %c1 : tensor<?x?x?xf32> | |
%2 = linalg.init_tensor [%0, %1, 1] : tensor<?x?x1xi64> | |
%3 = linalg.fill ins(%c0_i64 : i64) outs(%2 : tensor<?x?x1xi64>) -> tensor<?x?x1xi64> | |
%4 = linalg.init_tensor [%0, %1, 1] : tensor<?x?x1xf32> | |
%5 = linalg.fill ins(%cst_1 : f32) outs(%4 : tensor<?x?x1xf32>) -> tensor<?x?x1xf32> | |
%6:2 = linalg.generic {indexing_maps = [#map0, #map1, #map1], iterator_types = ["parallel", "parallel", "reduction"]} ins(%arg0 : tensor<?x?x?xf32>) outs(%5, %3 : tensor<?x?x1xf32>, tensor<?x?x1xi64>) { | |
^bb0(%arg1: f32, %arg2: f32, %arg3: i64): | |
%16 = linalg.index 2 : index | |
%17 = arith.index_cast %16 : index to i64 | |
%18 = arith.cmpf ogt, %arg1, %arg2 : f32 | |
%19 = arith.select %18, %arg1, %arg2 : f32 | |
%20 = arith.select %18, %17, %arg3 : i64 | |
linalg.yield %19, %20 : f32, i64 | |
} -> (tensor<?x?x1xf32>, tensor<?x?x1xi64>) | |
%7 = tensor.dim %arg0, %c2 : tensor<?x?x?xf32> | |
%8 = arith.cmpi eq, %0, %0 : index | |
cf.assert %8, "mismatched size for broadcast" | |
%9 = arith.cmpi eq, %1, %1 : index | |
cf.assert %9, "mismatched size for broadcast" | |
%10 = linalg.init_tensor [%0, %1, %7] : tensor<?x?x?xf32> | |
%11 = linalg.generic {indexing_maps = [#map0, #map1, #map0], iterator_types = ["parallel", "parallel", "parallel"]} ins(%arg0, %6#0 : tensor<?x?x?xf32>, tensor<?x?x1xf32>) outs(%10 : tensor<?x?x?xf32>) { | |
^bb0(%arg1: f32, %arg2: f32, %arg3: f32): | |
%16 = arith.truncf %cst_0 : f64 to f32 | |
%17 = arith.mulf %arg2, %16 : f32 | |
%18 = arith.subf %arg1, %17 : f32 | |
linalg.yield %18 : f32 | |
} -> tensor<?x?x?xf32> | |
%12 = linalg.generic {indexing_maps = [#map0, #map0], iterator_types = ["parallel", "parallel", "parallel"]} ins(%11 : tensor<?x?x?xf32>) outs(%10 : tensor<?x?x?xf32>) { | |
^bb0(%arg1: f32, %arg2: f32): | |
%16 = math.exp %arg1 : f32 | |
linalg.yield %16 : f32 | |
} -> tensor<?x?x?xf32> | |
%13 = linalg.fill ins(%cst : f32) outs(%4 : tensor<?x?x1xf32>) -> tensor<?x?x1xf32> | |
%14 = linalg.generic {indexing_maps = [#map0, #map1], iterator_types = ["parallel", "parallel", "reduction"]} ins(%12 : tensor<?x?x?xf32>) outs(%13 : tensor<?x?x1xf32>) { | |
^bb0(%arg1: f32, %arg2: f32): | |
%16 = arith.addf %arg1, %arg2 : f32 | |
linalg.yield %16 : f32 | |
} -> tensor<?x?x1xf32> | |
cf.assert %8, "mismatched size for broadcast" | |
cf.assert %9, "mismatched size for broadcast" | |
%15 = linalg.generic {indexing_maps = [#map0, #map1, #map0], iterator_types = ["parallel", "parallel", "parallel"]} ins(%12, %14 : tensor<?x?x?xf32>, tensor<?x?x1xf32>) outs(%10 : tensor<?x?x?xf32>) { | |
^bb0(%arg1: f32, %arg2: f32, %arg3: f32): | |
%16 = arith.divf %arg1, %arg2 : f32 | |
linalg.yield %16 : f32 | |
} -> tensor<?x?x?xf32> | |
return %15 : tensor<?x?x?xf32> | |
} | |
} |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment