llvm
/
torch-mlir
mirror of https://github.com/llvm/torch-mlir
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
torch-mlir/test/Conversion/TorchToStablehlo/view_like.mlir

575 lines
40 KiB
MLIR
Raw Normal View History

mhlo: migrate conversion to stablehlo (#1840) This patch replaces all MHLO operations with their StableHLO counterparts and adds a validation pass to ensure that no MHLO operations remain before translating all Stablehlo operations to the MHLO dialect for further lowering to the Linalg dialect. This patch also updates all lit tests so that they refer to the `convert-torch-to-stablehlo` pass and so that they check for StableHLO operations.
2023-02-02 21:29:47 +08:00
// RUN: torch-mlir-opt <%s -convert-torch-to-stablehlo -split-input-file -verify-diagnostics | FileCheck %s
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
// CHECK-LABEL: func.func @torch.aten.slice.strided$slice_like(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?,?],f32>) -> !torch.vtensor<[?,?,?],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?,?],f32> -> tensor<?x?x?xf32>
// CHECK: %[[INT0:.*]] = torch.constant.int 0
// CHECK: %[[T1:.*]] = torch_c.to_i64 %[[INT0]]
// CHECK: %[[INT2:.*]] = torch.constant.int 2
// CHECK: %[[T2:.*]] = torch_c.to_i64 %[[INT2]]
Fixes for split tensor and slice (#2314) * RecomposeComplexOps: Remove dead slice op * lib/Dialect/Torch/IR/TorchOps.cpp: Fold slice ops even when they are on non-value tensors * lib/Conversion/TorchToTosa/TorchToTosa.cpp: Fix slice start/end out of range/none * lib/Dialect/Torch/IR/TorchOps.cpp: AtenSliceTensorOp::fold: Fold slices that go from 0:int_max * More tests for aten.split.Tensor
2023-07-20 15:53:54 +08:00
// CHECK: %[[INT10:.*]] = torch.constant.int 10
// CHECK: %[[T3:.*]] = torch_c.to_i64 %[[INT10]]
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
// CHECK: %[[C0:.*]] = arith.constant 0 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM:.*]] = tensor.dim %[[T0]], %[[C0]] : tensor<?x?x?xf32>
// CHECK: %[[T4:.*]] = arith.index_cast %[[DIM]] : index to i64
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
// CHECK: %[[C0_I64:.*]] = arith.constant 0 : i64
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[T5:.*]] = arith.subi %[[C0_I64]], %[[T4]] : i64
// CHECK: %[[T6:.*]] = arith.maxsi %[[T5]], %[[T1]] : i64
// CHECK: %[[T7:.*]] = arith.minsi %[[T4]], %[[T6]] : i64
// CHECK: %[[T8:.*]] = arith.addi %[[T4]], %[[T7]] : i64
// CHECK: %[[T9:.*]] = arith.cmpi sge, %[[T7]], %[[C0_I64]] : i64
// CHECK: %[[T10:.*]] = arith.select %[[T9]], %[[T7]], %[[T8]] : i64
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
// CHECK: %[[C0_I64_0:.*]] = arith.constant 0 : i64
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[T11:.*]] = arith.subi %[[C0_I64_0]], %[[T4]] : i64
// CHECK: %[[T12:.*]] = arith.maxsi %[[T11]], %[[T3]] : i64
// CHECK: %[[T13:.*]] = arith.minsi %[[T4]], %[[T12]] : i64
// CHECK: %[[T14:.*]] = arith.addi %[[T4]], %[[T13]] : i64
// CHECK: %[[T15:.*]] = arith.cmpi sge, %[[T13]], %[[C0_I64_0]] : i64
// CHECK: %[[T16:.*]] = arith.select %[[T15]], %[[T13]], %[[T14]] : i64
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
// CHECK: %[[C0_1:.*]] = arith.constant 0 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM_2:.*]] = tensor.dim %[[T0]], %[[C0_1]] : tensor<?x?x?xf32>
// CHECK: %[[T17:.*]] = arith.index_cast %[[DIM_2]] : index to i64
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
// CHECK: %[[C1:.*]] = arith.constant 1 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM_3:.*]] = tensor.dim %[[T0]], %[[C1]] : tensor<?x?x?xf32>
// CHECK: %[[T18:.*]] = arith.index_cast %[[DIM_3]] : index to i64
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
// CHECK: %[[C2:.*]] = arith.constant 2 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM_4:.*]] = tensor.dim %[[T0]], %[[C2]] : tensor<?x?x?xf32>
// CHECK: %[[T19:.*]] = arith.index_cast %[[DIM_4]] : index to i64
// CHECK: %[[C0_I64_5:.*]] = arith.constant 0 : i64
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
// CHECK: %[[C1_I64:.*]] = arith.constant 1 : i64
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[T20:.*]] = arith.cmpi eq, %[[T16]], %[[C0_I64_5]] : i64
// CHECK: %[[T21:.*]] = arith.select %[[T20]], %[[T17]], %[[T16]] : i64
// CHECK: %[[FROM_ELEMENTS:.*]] = tensor.from_elements %[[T10]], %[[C0_I64_5]], %[[C0_I64_5]] : tensor<3xi64>
// CHECK: %[[FROM_ELEMENTS_6:.*]] = tensor.from_elements %[[T21]], %[[T18]], %[[T19]] : tensor<3xi64>
// CHECK: %[[FROM_ELEMENTS_7:.*]] = tensor.from_elements %[[T2]], %[[C1_I64]], %[[C1_I64]] : tensor<3xi64>
mhlo: migrate conversion to stablehlo (#1840) This patch replaces all MHLO operations with their StableHLO counterparts and adds a validation pass to ensure that no MHLO operations remain before translating all Stablehlo operations to the MHLO dialect for further lowering to the Linalg dialect. This patch also updates all lit tests so that they refer to the `convert-torch-to-stablehlo` pass and so that they check for StableHLO operations.
2023-02-02 21:29:47 +08:00
// CHECK: %[[T22:.*]] = stablehlo.real_dynamic_slice %[[T0]], %[[FROM_ELEMENTS]], %[[FROM_ELEMENTS]]_6, %[[FROM_ELEMENTS]]_7 : (tensor<?x?x?xf32>, tensor<3xi64>, tensor<3xi64>, tensor<3xi64>) -> tensor<?x?x?xf32>
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[T23:.*]] = torch_c.from_builtin_tensor %[[T22]] : tensor<?x?x?xf32> -> !torch.vtensor<[?,?,?],f32>
// CHECK: return %[[T23]] : !torch.vtensor<[?,?,?],f32>
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
func.func @torch.aten.slice.strided$slice_like(%arg0: !torch.vtensor<[?,?,?],f32>) -> !torch.vtensor<[?,?,?],f32> {
%int0 = torch.constant.int 0
%int2 = torch.constant.int 2
Fixes for split tensor and slice (#2314) * RecomposeComplexOps: Remove dead slice op * lib/Dialect/Torch/IR/TorchOps.cpp: Fold slice ops even when they are on non-value tensors * lib/Conversion/TorchToTosa/TorchToTosa.cpp: Fix slice start/end out of range/none * lib/Dialect/Torch/IR/TorchOps.cpp: AtenSliceTensorOp::fold: Fold slices that go from 0:int_max * More tests for aten.split.Tensor
2023-07-20 15:53:54 +08:00
%int10 = torch.constant.int 10
%0 = torch.aten.slice.Tensor %arg0, %int0, %int0, %int10, %int2 : !torch.vtensor<[?,?,?],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[?,?,?],f32>
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
return %0 : !torch.vtensor<[?,?,?],f32>
}
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// -----
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
// CHECK-LABEL: func.func @torch.aten.slice.strided.static$slice_like(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[4,65,256],f32>) -> !torch.vtensor<[2,65,256],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[4,65,256],f32> -> tensor<4x65x256xf32>
// CHECK: %[[INT0:.*]] = torch.constant.int 0
// CHECK: %[[T1:.*]] = torch_c.to_i64 %[[INT0]]
// CHECK: %[[INT2:.*]] = torch.constant.int 2
// CHECK: %[[T2:.*]] = torch_c.to_i64 %[[INT2]]
// CHECK: %[[INT9223372036854775807:.*]] = torch.constant.int 9223372036854775807
// CHECK: %[[T3:.*]] = torch_c.to_i64 %[[INT9223372036854775807]]
// CHECK: %[[C0:.*]] = arith.constant 0 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM:.*]] = tensor.dim %[[T0]], %[[C0]] : tensor<4x65x256xf32>
// CHECK: %[[T4:.*]] = arith.index_cast %[[DIM]] : index to i64
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
// CHECK: %[[C0_I64:.*]] = arith.constant 0 : i64
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[T5:.*]] = arith.subi %[[C0_I64]], %[[T4]] : i64
// CHECK: %[[T6:.*]] = arith.maxsi %[[T5]], %[[T1]] : i64
// CHECK: %[[T7:.*]] = arith.minsi %[[T4]], %[[T6]] : i64
// CHECK: %[[T8:.*]] = arith.addi %[[T4]], %[[T7]] : i64
// CHECK: %[[T9:.*]] = arith.cmpi sge, %[[T7]], %[[C0_I64]] : i64
// CHECK: %[[T10:.*]] = arith.select %[[T9]], %[[T7]], %[[T8]] : i64
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
// CHECK: %[[C0_I64_0:.*]] = arith.constant 0 : i64
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[T11:.*]] = arith.subi %[[C0_I64_0]], %[[T4]] : i64
// CHECK: %[[T12:.*]] = arith.maxsi %[[T11]], %[[T3]] : i64
// CHECK: %[[T13:.*]] = arith.minsi %[[T4]], %[[T12]] : i64
// CHECK: %[[T14:.*]] = arith.addi %[[T4]], %[[T13]] : i64
// CHECK: %[[T15:.*]] = arith.cmpi sge, %[[T13]], %[[C0_I64_0]] : i64
// CHECK: %[[T16:.*]] = arith.select %[[T15]], %[[T13]], %[[T14]] : i64
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
// CHECK: %[[C0_1:.*]] = arith.constant 0 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM_2:.*]] = tensor.dim %[[T0]], %[[C0_1]] : tensor<4x65x256xf32>
// CHECK: %[[T17:.*]] = arith.index_cast %[[DIM_2]] : index to i64
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
// CHECK: %[[C1:.*]] = arith.constant 1 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM_3:.*]] = tensor.dim %[[T0]], %[[C1]] : tensor<4x65x256xf32>
// CHECK: %[[T18:.*]] = arith.index_cast %[[DIM_3]] : index to i64
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
// CHECK: %[[C2:.*]] = arith.constant 2 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM_4:.*]] = tensor.dim %[[T0]], %[[C2]] : tensor<4x65x256xf32>
// CHECK: %[[T19:.*]] = arith.index_cast %[[DIM_4]] : index to i64
// CHECK: %[[C0_I64_5:.*]] = arith.constant 0 : i64
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
// CHECK: %[[C1_I64:.*]] = arith.constant 1 : i64
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[T20:.*]] = arith.cmpi eq, %[[T16]], %[[C0_I64_5]] : i64
// CHECK: %[[T21:.*]] = arith.select %[[T20]], %[[T17]], %[[T16]] : i64
// CHECK: %[[FROM_ELEMENTS:.*]] = tensor.from_elements %[[T10]], %[[C0_I64_5]], %[[C0_I64_5]] : tensor<3xi64>
// CHECK: %[[FROM_ELEMENTS_6:.*]] = tensor.from_elements %[[T21]], %[[T18]], %[[T19]] : tensor<3xi64>
// CHECK: %[[FROM_ELEMENTS_7:.*]] = tensor.from_elements %[[T2]], %[[C1_I64]], %[[C1_I64]] : tensor<3xi64>
mhlo: migrate conversion to stablehlo (#1840) This patch replaces all MHLO operations with their StableHLO counterparts and adds a validation pass to ensure that no MHLO operations remain before translating all Stablehlo operations to the MHLO dialect for further lowering to the Linalg dialect. This patch also updates all lit tests so that they refer to the `convert-torch-to-stablehlo` pass and so that they check for StableHLO operations.
2023-02-02 21:29:47 +08:00
// CHECK: %[[T22:.*]] = stablehlo.real_dynamic_slice %[[T0]], %[[FROM_ELEMENTS]], %[[FROM_ELEMENTS]]_6, %[[FROM_ELEMENTS]]_7 : (tensor<4x65x256xf32>, tensor<3xi64>, tensor<3xi64>, tensor<3xi64>) -> tensor<2x65x256xf32>
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[T23:.*]] = torch_c.from_builtin_tensor %[[T22]] : tensor<2x65x256xf32> -> !torch.vtensor<[2,65,256],f32>
// CHECK: return %[[T23]] : !torch.vtensor<[2,65,256],f32>
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
func.func @torch.aten.slice.strided.static$slice_like(%arg0: !torch.vtensor<[4,65,256],f32>) -> !torch.vtensor<[2,65,256],f32> {
%int0 = torch.constant.int 0
%int2 = torch.constant.int 2
%int9223372036854775807 = torch.constant.int 9223372036854775807
%0 = torch.aten.slice.Tensor %arg0, %int0, %int0, %int9223372036854775807, %int2 : !torch.vtensor<[4,65,256],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[2,65,256],f32>
return %0 : !torch.vtensor<[2,65,256],f32>
}
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// -----
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
// CHECK-LABEL: func.func @torch.aten.slice.last$slice_like(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?,?],f32>) -> !torch.vtensor<[?,1,?],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?,?],f32> -> tensor<?x?x?xf32>
// CHECK: %[[INT0:.*]] = torch.constant.int 0
// CHECK: %[[T1:.*]] = torch_c.to_i64 %[[INT0]]
// CHECK: %[[INT1:.*]] = torch.constant.int 1
// CHECK: %[[T2:.*]] = torch_c.to_i64 %[[INT1]]
// CHECK: %[[INT:.*]]-1 = torch.constant.int -1
// CHECK: %[[T3:.*]] = torch_c.to_i64 %[[INT]]-1
// CHECK: %[[C1:.*]] = arith.constant 1 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM:.*]] = tensor.dim %[[T0]], %[[C1]] : tensor<?x?x?xf32>
// CHECK: %[[T4:.*]] = arith.index_cast %[[DIM]] : index to i64
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
// CHECK: %[[C0_I64:.*]] = arith.constant 0 : i64
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[T5:.*]] = arith.subi %[[C0_I64]], %[[T4]] : i64
// CHECK: %[[T6:.*]] = arith.maxsi %[[T5]], %[[T3]] : i64
// CHECK: %[[T7:.*]] = arith.minsi %[[T4]], %[[T6]] : i64
// CHECK: %[[T8:.*]] = arith.addi %[[T4]], %[[T7]] : i64
// CHECK: %[[T9:.*]] = arith.cmpi sge, %[[T7]], %[[C0_I64]] : i64
// CHECK: %[[T10:.*]] = arith.select %[[T9]], %[[T7]], %[[T8]] : i64
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
// CHECK: %[[C0_I64_0:.*]] = arith.constant 0 : i64
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[T11:.*]] = arith.subi %[[C0_I64_0]], %[[T4]] : i64
// CHECK: %[[T12:.*]] = arith.maxsi %[[T11]], %[[T1]] : i64
// CHECK: %[[T13:.*]] = arith.minsi %[[T4]], %[[T12]] : i64
// CHECK: %[[T14:.*]] = arith.addi %[[T4]], %[[T13]] : i64
// CHECK: %[[T15:.*]] = arith.cmpi sge, %[[T13]], %[[C0_I64_0]] : i64
// CHECK: %[[T16:.*]] = arith.select %[[T15]], %[[T13]], %[[T14]] : i64
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
// CHECK: %[[C0:.*]] = arith.constant 0 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM_1:.*]] = tensor.dim %[[T0]], %[[C0]] : tensor<?x?x?xf32>
// CHECK: %[[T17:.*]] = arith.index_cast %[[DIM_1]] : index to i64
// CHECK: %[[C1_2:.*]] = arith.constant 1 : index
// CHECK: %[[DIM_3:.*]] = tensor.dim %[[T0]], %[[C1_2]] : tensor<?x?x?xf32>
// CHECK: %[[T18:.*]] = arith.index_cast %[[DIM_3]] : index to i64
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
// CHECK: %[[C2:.*]] = arith.constant 2 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM_4:.*]] = tensor.dim %[[T0]], %[[C2]] : tensor<?x?x?xf32>
// CHECK: %[[T19:.*]] = arith.index_cast %[[DIM_4]] : index to i64
// CHECK: %[[C0_I64_5:.*]] = arith.constant 0 : i64
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
// CHECK: %[[C1_I64:.*]] = arith.constant 1 : i64
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[T20:.*]] = arith.cmpi eq, %[[T16]], %[[C0_I64_5]] : i64
// CHECK: %[[T21:.*]] = arith.select %[[T20]], %[[T18]], %[[T16]] : i64
// CHECK: %[[FROM_ELEMENTS:.*]] = tensor.from_elements %[[C0_I64_5]], %[[T10]], %[[C0_I64_5]] : tensor<3xi64>
// CHECK: %[[FROM_ELEMENTS_6:.*]] = tensor.from_elements %[[T17]], %[[T21]], %[[T19]] : tensor<3xi64>
// CHECK: %[[FROM_ELEMENTS_7:.*]] = tensor.from_elements %[[C1_I64]], %[[T2]], %[[C1_I64]] : tensor<3xi64>
mhlo: migrate conversion to stablehlo (#1840) This patch replaces all MHLO operations with their StableHLO counterparts and adds a validation pass to ensure that no MHLO operations remain before translating all Stablehlo operations to the MHLO dialect for further lowering to the Linalg dialect. This patch also updates all lit tests so that they refer to the `convert-torch-to-stablehlo` pass and so that they check for StableHLO operations.
2023-02-02 21:29:47 +08:00
// CHECK: %[[T22:.*]] = stablehlo.real_dynamic_slice %[[T0]], %[[FROM_ELEMENTS]], %[[FROM_ELEMENTS]]_6, %[[FROM_ELEMENTS]]_7 : (tensor<?x?x?xf32>, tensor<3xi64>, tensor<3xi64>, tensor<3xi64>) -> tensor<?x1x?xf32>
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[T23:.*]] = torch_c.from_builtin_tensor %[[T22]] : tensor<?x1x?xf32> -> !torch.vtensor<[?,1,?],f32>
// CHECK: return %[[T23]] : !torch.vtensor<[?,1,?],f32>
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
func.func @torch.aten.slice.last$slice_like(%arg0: !torch.vtensor<[?,?,?],f32>) -> !torch.vtensor<[?,1,?],f32> {
%int0 = torch.constant.int 0
%int1 = torch.constant.int 1
%int-1 = torch.constant.int -1
%0 = torch.aten.slice.Tensor %arg0, %int1, %int-1, %int0, %int1 : !torch.vtensor<[?,?,?],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[?,1,?],f32>
return %0 : !torch.vtensor<[?,1,?],f32>
}
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// -----
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
// CHECK-LABEL: func.func @torch.aten.slice.last.static$slice_like(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[4,65,256],f32>) -> !torch.vtensor<[4,1,256],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[4,65,256],f32> -> tensor<4x65x256xf32>
// CHECK: %[[INT0:.*]] = torch.constant.int 0
// CHECK: %[[T1:.*]] = torch_c.to_i64 %[[INT0]]
// CHECK: %[[INT1:.*]] = torch.constant.int 1
// CHECK: %[[T2:.*]] = torch_c.to_i64 %[[INT1]]
// CHECK: %[[INT:.*]]-1 = torch.constant.int -1
// CHECK: %[[T3:.*]] = torch_c.to_i64 %[[INT]]-1
// CHECK: %[[C1:.*]] = arith.constant 1 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM:.*]] = tensor.dim %[[T0]], %[[C1]] : tensor<4x65x256xf32>
// CHECK: %[[T4:.*]] = arith.index_cast %[[DIM]] : index to i64
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
// CHECK: %[[C0_I64:.*]] = arith.constant 0 : i64
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[T5:.*]] = arith.subi %[[C0_I64]], %[[T4]] : i64
// CHECK: %[[T6:.*]] = arith.maxsi %[[T5]], %[[T3]] : i64
// CHECK: %[[T7:.*]] = arith.minsi %[[T4]], %[[T6]] : i64
// CHECK: %[[T8:.*]] = arith.addi %[[T4]], %[[T7]] : i64
// CHECK: %[[T9:.*]] = arith.cmpi sge, %[[T7]], %[[C0_I64]] : i64
// CHECK: %[[T10:.*]] = arith.select %[[T9]], %[[T7]], %[[T8]] : i64
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
// CHECK: %[[C0_I64_0:.*]] = arith.constant 0 : i64
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[T11:.*]] = arith.subi %[[C0_I64_0]], %[[T4]] : i64
// CHECK: %[[T12:.*]] = arith.maxsi %[[T11]], %[[T1]] : i64
// CHECK: %[[T13:.*]] = arith.minsi %[[T4]], %[[T12]] : i64
// CHECK: %[[T14:.*]] = arith.addi %[[T4]], %[[T13]] : i64
// CHECK: %[[T15:.*]] = arith.cmpi sge, %[[T13]], %[[C0_I64_0]] : i64
// CHECK: %[[T16:.*]] = arith.select %[[T15]], %[[T13]], %[[T14]] : i64
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
// CHECK: %[[C0:.*]] = arith.constant 0 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM_1:.*]] = tensor.dim %[[T0]], %[[C0]] : tensor<4x65x256xf32>
// CHECK: %[[T17:.*]] = arith.index_cast %[[DIM_1]] : index to i64
// CHECK: %[[C1_2:.*]] = arith.constant 1 : index
// CHECK: %[[DIM_3:.*]] = tensor.dim %[[T0]], %[[C1_2]] : tensor<4x65x256xf32>
// CHECK: %[[T18:.*]] = arith.index_cast %[[DIM_3]] : index to i64
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
// CHECK: %[[C2:.*]] = arith.constant 2 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM_4:.*]] = tensor.dim %[[T0]], %[[C2]] : tensor<4x65x256xf32>
// CHECK: %[[T19:.*]] = arith.index_cast %[[DIM_4]] : index to i64
// CHECK: %[[C0_I64_5:.*]] = arith.constant 0 : i64
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
// CHECK: %[[C1_I64:.*]] = arith.constant 1 : i64
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[T20:.*]] = arith.cmpi eq, %[[T16]], %[[C0_I64_5]] : i64
// CHECK: %[[T21:.*]] = arith.select %[[T20]], %[[T18]], %[[T16]] : i64
// CHECK: %[[FROM_ELEMENTS:.*]] = tensor.from_elements %[[C0_I64_5]], %[[T10]], %[[C0_I64_5]] : tensor<3xi64>
// CHECK: %[[FROM_ELEMENTS_6:.*]] = tensor.from_elements %[[T17]], %[[T21]], %[[T19]] : tensor<3xi64>
// CHECK: %[[FROM_ELEMENTS_7:.*]] = tensor.from_elements %[[C1_I64]], %[[T2]], %[[C1_I64]] : tensor<3xi64>
mhlo: migrate conversion to stablehlo (#1840) This patch replaces all MHLO operations with their StableHLO counterparts and adds a validation pass to ensure that no MHLO operations remain before translating all Stablehlo operations to the MHLO dialect for further lowering to the Linalg dialect. This patch also updates all lit tests so that they refer to the `convert-torch-to-stablehlo` pass and so that they check for StableHLO operations.
2023-02-02 21:29:47 +08:00
// CHECK: %[[T22:.*]] = stablehlo.real_dynamic_slice %[[T0]], %[[FROM_ELEMENTS]], %[[FROM_ELEMENTS]]_6, %[[FROM_ELEMENTS]]_7 : (tensor<4x65x256xf32>, tensor<3xi64>, tensor<3xi64>, tensor<3xi64>) -> tensor<4x1x256xf32>
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[T23:.*]] = torch_c.from_builtin_tensor %[[T22]] : tensor<4x1x256xf32> -> !torch.vtensor<[4,1,256],f32>
// CHECK: return %[[T23]] : !torch.vtensor<[4,1,256],f32>
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
func.func @torch.aten.slice.last.static$slice_like(%arg0: !torch.vtensor<[4,65,256],f32>) -> !torch.vtensor<[4,1,256],f32> {
%int0 = torch.constant.int 0
%int1 = torch.constant.int 1
%int-1 = torch.constant.int -1
%0 = torch.aten.slice.Tensor %arg0, %int1, %int-1, %int0, %int1 : !torch.vtensor<[4,65,256],f32>, !torch.int, !torch.int, !torch.int, !torch.int -> !torch.vtensor<[4,1,256],f32>
return %0 : !torch.vtensor<[4,1,256],f32>
}
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// -----
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
// CHECK-LABEL: func.func @torch.aten.slice.none$slice_like(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?,?],f32>) -> !torch.vtensor<[?,?,?],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?,?],f32> -> tensor<?x?x?xf32>
// CHECK: %[[INT1:.*]] = torch.constant.int 1
// CHECK: %[[INT2:.*]] = torch.constant.int 2
// CHECK: %[[T1:.*]] = torch_c.to_i64 %[[INT2]]
// CHECK: %[[NONE:.*]] = torch.constant.none
// CHECK: %[[C1:.*]] = arith.constant 1 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM:.*]] = tensor.dim %[[T0]], %[[C1]] : tensor<?x?x?xf32>
// CHECK: %[[T2:.*]] = arith.index_cast %[[DIM]] : index to i64
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
// CHECK: %[[C0_I64:.*]] = arith.constant 0 : i64
// CHECK: %[[C0:.*]] = arith.constant 0 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM_0:.*]] = tensor.dim %[[T0]], %[[C0]] : tensor<?x?x?xf32>
// CHECK: %[[T3:.*]] = arith.index_cast %[[DIM_0]] : index to i64
// CHECK: %[[C1_1:.*]] = arith.constant 1 : index
// CHECK: %[[DIM_2:.*]] = tensor.dim %[[T0]], %[[C1_1]] : tensor<?x?x?xf32>
// CHECK: %[[T4:.*]] = arith.index_cast %[[DIM_2]] : index to i64
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
// CHECK: %[[C2:.*]] = arith.constant 2 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM_3:.*]] = tensor.dim %[[T0]], %[[C2]] : tensor<?x?x?xf32>
// CHECK: %[[T5:.*]] = arith.index_cast %[[DIM_3]] : index to i64
// CHECK: %[[C0_I64_4:.*]] = arith.constant 0 : i64
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
// CHECK: %[[C1_I64:.*]] = arith.constant 1 : i64
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[T6:.*]] = arith.cmpi eq, %[[T2]], %[[C0_I64_4]] : i64
// CHECK: %[[T7:.*]] = arith.select %[[T6]], %[[T4]], %[[T2]] : i64
// CHECK: %[[FROM_ELEMENTS:.*]] = tensor.from_elements %[[C0_I64_4]], %[[C0_I64]], %[[C0_I64_4]] : tensor<3xi64>
// CHECK: %[[FROM_ELEMENTS_5:.*]] = tensor.from_elements %[[T3]], %[[T7]], %[[T5]] : tensor<3xi64>
// CHECK: %[[FROM_ELEMENTS_6:.*]] = tensor.from_elements %[[C1_I64]], %[[T1]], %[[C1_I64]] : tensor<3xi64>
mhlo: migrate conversion to stablehlo (#1840) This patch replaces all MHLO operations with their StableHLO counterparts and adds a validation pass to ensure that no MHLO operations remain before translating all Stablehlo operations to the MHLO dialect for further lowering to the Linalg dialect. This patch also updates all lit tests so that they refer to the `convert-torch-to-stablehlo` pass and so that they check for StableHLO operations.
2023-02-02 21:29:47 +08:00
// CHECK: %[[T8:.*]] = stablehlo.real_dynamic_slice %[[T0]], %[[FROM_ELEMENTS]], %[[FROM_ELEMENTS]]_5, %[[FROM_ELEMENTS]]_6 : (tensor<?x?x?xf32>, tensor<3xi64>, tensor<3xi64>, tensor<3xi64>) -> tensor<?x?x?xf32>
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[T9:.*]] = torch_c.from_builtin_tensor %[[T8]] : tensor<?x?x?xf32> -> !torch.vtensor<[?,?,?],f32>
// CHECK: return %[[T9]] : !torch.vtensor<[?,?,?],f32>
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
func.func @torch.aten.slice.none$slice_like(%arg0: !torch.vtensor<[?,?,?],f32>) -> !torch.vtensor<[?,?,?],f32> {
%int1 = torch.constant.int 1
%int2 = torch.constant.int 2
%none = torch.constant.none
%0 = torch.aten.slice.Tensor %arg0, %int1, %none, %none, %int2 : !torch.vtensor<[?,?,?],f32>, !torch.int, !torch.none, !torch.none, !torch.int -> !torch.vtensor<[?,?,?],f32>
return %0 : !torch.vtensor<[?,?,?],f32>
}
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// -----
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
// CHECK-LABEL: func.func @torch.aten.slice.none.static$slice_like(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[4,65,256],f32>) -> !torch.vtensor<[4,33,256],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[4,65,256],f32> -> tensor<4x65x256xf32>
// CHECK: %[[INT1:.*]] = torch.constant.int 1
// CHECK: %[[INT2:.*]] = torch.constant.int 2
// CHECK: %[[T1:.*]] = torch_c.to_i64 %[[INT2]]
// CHECK: %[[NONE:.*]] = torch.constant.none
// CHECK: %[[C1:.*]] = arith.constant 1 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM:.*]] = tensor.dim %[[T0]], %[[C1]] : tensor<4x65x256xf32>
// CHECK: %[[T2:.*]] = arith.index_cast %[[DIM]] : index to i64
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
// CHECK: %[[C0_I64:.*]] = arith.constant 0 : i64
// CHECK: %[[C0:.*]] = arith.constant 0 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM_0:.*]] = tensor.dim %[[T0]], %[[C0]] : tensor<4x65x256xf32>
// CHECK: %[[T3:.*]] = arith.index_cast %[[DIM_0]] : index to i64
// CHECK: %[[C1_1:.*]] = arith.constant 1 : index
// CHECK: %[[DIM_2:.*]] = tensor.dim %[[T0]], %[[C1_1]] : tensor<4x65x256xf32>
// CHECK: %[[T4:.*]] = arith.index_cast %[[DIM_2]] : index to i64
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
// CHECK: %[[C2:.*]] = arith.constant 2 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM_3:.*]] = tensor.dim %[[T0]], %[[C2]] : tensor<4x65x256xf32>
// CHECK: %[[T5:.*]] = arith.index_cast %[[DIM_3]] : index to i64
// CHECK: %[[C0_I64_4:.*]] = arith.constant 0 : i64
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
// CHECK: %[[C1_I64:.*]] = arith.constant 1 : i64
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[T6:.*]] = arith.cmpi eq, %[[T2]], %[[C0_I64_4]] : i64
// CHECK: %[[T7:.*]] = arith.select %[[T6]], %[[T4]], %[[T2]] : i64
// CHECK: %[[FROM_ELEMENTS:.*]] = tensor.from_elements %[[C0_I64_4]], %[[C0_I64]], %[[C0_I64_4]] : tensor<3xi64>
// CHECK: %[[FROM_ELEMENTS_5:.*]] = tensor.from_elements %[[T3]], %[[T7]], %[[T5]] : tensor<3xi64>
// CHECK: %[[FROM_ELEMENTS_6:.*]] = tensor.from_elements %[[C1_I64]], %[[T1]], %[[C1_I64]] : tensor<3xi64>
mhlo: migrate conversion to stablehlo (#1840) This patch replaces all MHLO operations with their StableHLO counterparts and adds a validation pass to ensure that no MHLO operations remain before translating all Stablehlo operations to the MHLO dialect for further lowering to the Linalg dialect. This patch also updates all lit tests so that they refer to the `convert-torch-to-stablehlo` pass and so that they check for StableHLO operations.
2023-02-02 21:29:47 +08:00
// CHECK: %[[T8:.*]] = stablehlo.real_dynamic_slice %[[T0]], %[[FROM_ELEMENTS]], %[[FROM_ELEMENTS]]_5, %[[FROM_ELEMENTS]]_6 : (tensor<4x65x256xf32>, tensor<3xi64>, tensor<3xi64>, tensor<3xi64>) -> tensor<4x33x256xf32>
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[T9:.*]] = torch_c.from_builtin_tensor %[[T8]] : tensor<4x33x256xf32> -> !torch.vtensor<[4,33,256],f32>
// CHECK: return %[[T9]] : !torch.vtensor<[4,33,256],f32>
[MHLO] Init MHLO slice like op patterns (#1091) See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-07-22 11:32:45 +08:00
func.func @torch.aten.slice.none.static$slice_like(%arg0: !torch.vtensor<[4,65,256],f32>) -> !torch.vtensor<[4,33,256],f32> {
%int1 = torch.constant.int 1
%int2 = torch.constant.int 2
%none = torch.constant.none
%0 = torch.aten.slice.Tensor %arg0, %int1, %none, %none, %int2 : !torch.vtensor<[4,65,256],f32>, !torch.int, !torch.none, !torch.none, !torch.int -> !torch.vtensor<[4,33,256],f32>
return %0 : !torch.vtensor<[4,33,256],f32>
}
[MHLO] Init MHLO view like op patterns (#1090) * [MHLO] Init MHLO view like op patterns See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com * update filecheck test cases * rebase, remove chlo and clang-format
2022-07-22 15:18:18 +08:00
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// -----
[MHLO] Init MHLO view like op patterns (#1090) * [MHLO] Init MHLO view like op patterns See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com * update filecheck test cases * rebase, remove chlo and clang-format
2022-07-22 15:18:18 +08:00
// CHECK-LABEL: func.func @torch.aten.view$basic(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?,?,?],f32>) -> !torch.vtensor<[?,224],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?,?,?],f32> -> tensor<?x?x?x?xf32>
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[INT:.*]]-1 = torch.constant.int -1
[MHLO] Init MHLO view like op patterns (#1090) * [MHLO] Init MHLO view like op patterns See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com * update filecheck test cases * rebase, remove chlo and clang-format
2022-07-22 15:18:18 +08:00
// CHECK: %[[INT224:.*]] = torch.constant.int 224
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[T1:.*]] = torch.prim.ListConstruct %[[INT]]-1, %[[INT]]224 : (!torch.int, !torch.int) -> !torch.list<int>
// CHECK: %[[T2:.*]] = torch_c.to_i64 %[[INT]]-1
[MHLO] Init MHLO view like op patterns (#1090) * [MHLO] Init MHLO view like op patterns See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com * update filecheck test cases * rebase, remove chlo and clang-format
2022-07-22 15:18:18 +08:00
// CHECK: %[[T3:.*]] = torch_c.to_i64 %[[INT224]]
// CHECK: %[[C1_I64:.*]] = arith.constant 1 : i64
// CHECK: %[[T4:.*]] = arith.muli %[[C1_I64]], %[[T2]] : i64
// CHECK: %[[T5:.*]] = arith.muli %[[T4]], %[[T3]] : i64
// CHECK: %[[T6:.*]] = arith.index_cast %[[T5]] : i64 to index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[FROM_ELEMENTS:.*]] = tensor.from_elements %[[T2]], %[[T3]] : tensor<2xi64>
mhlo: migrate conversion to stablehlo (#1840) This patch replaces all MHLO operations with their StableHLO counterparts and adds a validation pass to ensure that no MHLO operations remain before translating all Stablehlo operations to the MHLO dialect for further lowering to the Linalg dialect. This patch also updates all lit tests so that they refer to the `convert-torch-to-stablehlo` pass and so that they check for StableHLO operations.
2023-02-02 21:29:47 +08:00
// CHECK: %[[T7:.*]] = stablehlo.compute_reshape_shape %[[T6]], %[[FROM_ELEMENTS]] : (index, tensor<2xi64>) -> tensor<2xi64>
// CHECK: %[[T8:.*]] = stablehlo.dynamic_reshape %[[T0]], %[[T7]] : (tensor<?x?x?x?xf32>, tensor<2xi64>) -> tensor<?x224xf32>
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[T9:.*]] = torch_c.from_builtin_tensor %[[T8]] : tensor<?x224xf32> -> !torch.vtensor<[?,224],f32>
// CHECK: return %[[T9]] : !torch.vtensor<[?,224],f32>
[MHLO] Init MHLO view like op patterns (#1090) * [MHLO] Init MHLO view like op patterns See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com * update filecheck test cases * rebase, remove chlo and clang-format
2022-07-22 15:18:18 +08:00
func.func @torch.aten.view$basic(%arg0: !torch.vtensor<[?,?,?,?],f32>) -> !torch.vtensor<[?,224],f32> {
%int-1 = torch.constant.int -1
%int224 = torch.constant.int 224
%0 = torch.prim.ListConstruct %int-1, %int224 : (!torch.int, !torch.int) -> !torch.list<int>
%1 = torch.aten.view %arg0, %0 : !torch.vtensor<[?,?,?,?],f32>, !torch.list<int> -> !torch.vtensor<[?,224],f32>
return %1 : !torch.vtensor<[?,224],f32>
}
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// -----
[MHLO] Init MHLO view like op patterns (#1090) * [MHLO] Init MHLO view like op patterns See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com * update filecheck test cases * rebase, remove chlo and clang-format
2022-07-22 15:18:18 +08:00
// CHECK-LABEL: func.func @torch.aten.reshape$basic(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?,?,?,?],f32>) -> !torch.vtensor<[?,120,4,64],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?,?,?,?],f32> -> tensor<?x?x?x?x?xf32>
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[INT:.*]]-1 = torch.constant.int -1
[MHLO] Init MHLO view like op patterns (#1090) * [MHLO] Init MHLO view like op patterns See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com * update filecheck test cases * rebase, remove chlo and clang-format
2022-07-22 15:18:18 +08:00
// CHECK: %[[INT120:.*]] = torch.constant.int 120
// CHECK: %[[INT4:.*]] = torch.constant.int 4
// CHECK: %[[INT64:.*]] = torch.constant.int 64
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[T1:.*]] = torch.prim.ListConstruct %[[INT]]-1, %[[INT]]120, %[[INT]]4, %[[INT]]64 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
// CHECK: %[[T2:.*]] = torch_c.to_i64 %[[INT]]-1
[MHLO] Init MHLO view like op patterns (#1090) * [MHLO] Init MHLO view like op patterns See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com * update filecheck test cases * rebase, remove chlo and clang-format
2022-07-22 15:18:18 +08:00
// CHECK: %[[T3:.*]] = torch_c.to_i64 %[[INT120]]
// CHECK: %[[T4:.*]] = torch_c.to_i64 %[[INT4]]
// CHECK: %[[T5:.*]] = torch_c.to_i64 %[[INT64]]
// CHECK: %[[C1_I64:.*]] = arith.constant 1 : i64
// CHECK: %[[T6:.*]] = arith.muli %[[C1_I64]], %[[T2]] : i64
// CHECK: %[[T7:.*]] = arith.muli %[[T6]], %[[T3]] : i64
// CHECK: %[[T8:.*]] = arith.muli %[[T7]], %[[T4]] : i64
// CHECK: %[[T9:.*]] = arith.muli %[[T8]], %[[T5]] : i64
// CHECK: %[[T10:.*]] = arith.index_cast %[[T9]] : i64 to index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[FROM_ELEMENTS:.*]] = tensor.from_elements %[[T2]], %[[T3]], %[[T4]], %[[T5]] : tensor<4xi64>
mhlo: migrate conversion to stablehlo (#1840) This patch replaces all MHLO operations with their StableHLO counterparts and adds a validation pass to ensure that no MHLO operations remain before translating all Stablehlo operations to the MHLO dialect for further lowering to the Linalg dialect. This patch also updates all lit tests so that they refer to the `convert-torch-to-stablehlo` pass and so that they check for StableHLO operations.
2023-02-02 21:29:47 +08:00
// CHECK: %[[T11:.*]] = stablehlo.compute_reshape_shape %[[T10]], %[[FROM_ELEMENTS]] : (index, tensor<4xi64>) -> tensor<4xi64>
// CHECK: %[[T12:.*]] = stablehlo.dynamic_reshape %[[T0]], %[[T11]] : (tensor<?x?x?x?x?xf32>, tensor<4xi64>) -> tensor<?x120x4x64xf32>
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[T13:.*]] = torch_c.from_builtin_tensor %[[T12]] : tensor<?x120x4x64xf32> -> !torch.vtensor<[?,120,4,64],f32>
// CHECK: return %[[T13]] : !torch.vtensor<[?,120,4,64],f32>
[MHLO] Init MHLO view like op patterns (#1090) * [MHLO] Init MHLO view like op patterns See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com * update filecheck test cases * rebase, remove chlo and clang-format
2022-07-22 15:18:18 +08:00
func.func @torch.aten.reshape$basic(%arg0: !torch.vtensor<[?,?,?,?,?],f32>) -> !torch.vtensor<[?,120,4,64],f32> {
%int-1 = torch.constant.int -1
%int120 = torch.constant.int 120
%int4 = torch.constant.int 4
%int64 = torch.constant.int 64
%0 = torch.prim.ListConstruct %int-1, %int120, %int4, %int64 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
%1 = torch.aten.reshape %arg0, %0 : !torch.vtensor<[?,?,?,?,?],f32>, !torch.list<int> -> !torch.vtensor<[?,120,4,64],f32>
return %1 : !torch.vtensor<[?,120,4,64],f32>
}
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// -----
[MHLO] Init MHLO view like op patterns (#1090) * [MHLO] Init MHLO view like op patterns See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com * update filecheck test cases * rebase, remove chlo and clang-format
2022-07-22 15:18:18 +08:00
// CHECK-LABEL: func.func @torch.aten.view$to_rank1(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[],f32>) -> !torch.vtensor<[1],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[],f32> -> tensor<f32>
// CHECK: %[[INT1:.*]] = torch.constant.int 1
// CHECK: %[[T1:.*]] = torch.prim.ListConstruct %[[INT1]] : (!torch.int) -> !torch.list<int>
mhlo: migrate conversion to stablehlo (#1840) This patch replaces all MHLO operations with their StableHLO counterparts and adds a validation pass to ensure that no MHLO operations remain before translating all Stablehlo operations to the MHLO dialect for further lowering to the Linalg dialect. This patch also updates all lit tests so that they refer to the `convert-torch-to-stablehlo` pass and so that they check for StableHLO operations.
2023-02-02 21:29:47 +08:00
// CHECK: %[[T2:.*]] = stablehlo.reshape %[[T0]] : (tensor<f32>) -> tensor<1xf32>
[MHLO] Init MHLO view like op patterns (#1090) * [MHLO] Init MHLO view like op patterns See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com * update filecheck test cases * rebase, remove chlo and clang-format
2022-07-22 15:18:18 +08:00
// CHECK: %[[T3:.*]] = torch_c.from_builtin_tensor %[[T2]] : tensor<1xf32> -> !torch.vtensor<[1],f32>
// CHECK: return %[[T3]] : !torch.vtensor<[1],f32>
func.func @torch.aten.view$to_rank1(%arg0: !torch.vtensor<[],f32>) -> !torch.vtensor<[1],f32> {
%int1 = torch.constant.int 1
%0 = torch.prim.ListConstruct %int1 : (!torch.int) -> !torch.list<int>
%1 = torch.aten.view %arg0, %0 : !torch.vtensor<[],f32>, !torch.list<int> -> !torch.vtensor<[1],f32>
return %1 : !torch.vtensor<[1],f32>
}
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// -----
[MHLO] Init MHLO view like op patterns (#1090) * [MHLO] Init MHLO view like op patterns See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com * update filecheck test cases * rebase, remove chlo and clang-format
2022-07-22 15:18:18 +08:00
// CHECK-LABEL: func.func @torch.aten.view$to_rank0(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[1],f32>) -> !torch.vtensor<[],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[1],f32> -> tensor<1xf32>
// CHECK: %[[T1:.*]] = torch.prim.ListConstruct : () -> !torch.list<int>
mhlo: migrate conversion to stablehlo (#1840) This patch replaces all MHLO operations with their StableHLO counterparts and adds a validation pass to ensure that no MHLO operations remain before translating all Stablehlo operations to the MHLO dialect for further lowering to the Linalg dialect. This patch also updates all lit tests so that they refer to the `convert-torch-to-stablehlo` pass and so that they check for StableHLO operations.
2023-02-02 21:29:47 +08:00
// CHECK: %[[T2:.*]] = stablehlo.reshape %[[T0]] : (tensor<1xf32>) -> tensor<f32>
[MHLO] Init MHLO view like op patterns (#1090) * [MHLO] Init MHLO view like op patterns See RFC: https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com * update filecheck test cases * rebase, remove chlo and clang-format
2022-07-22 15:18:18 +08:00
// CHECK: %[[T3:.*]] = torch_c.from_builtin_tensor %[[T2]] : tensor<f32> -> !torch.vtensor<[],f32>
// CHECK: return %[[T3]] : !torch.vtensor<[],f32>
func.func @torch.aten.view$to_rank0(%arg0: !torch.vtensor<[1],f32>) -> !torch.vtensor<[],f32> {
%0 = torch.prim.ListConstruct : () -> !torch.list<int>
%1 = torch.aten.view %arg0, %0 : !torch.vtensor<[1],f32>, !torch.list<int> -> !torch.vtensor<[],f32>
return %1 : !torch.vtensor<[],f32>
}
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// -----
[MHLO] Add [un]squeeze op patterns (#1099) * [MHLO] Add [un]squeeze op patterns * Conform to llvm coding standard * minor update
2022-07-25 23:28:48 +08:00
// CHECK-LABEL: func.func @torch.aten.squeeze.dim$0$static(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[2,1,2,1,2],f32>) -> !torch.vtensor<[2,1,2,1,2],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[2,1,2,1,2],f32> -> tensor<2x1x2x1x2xf32>
// CHECK: %[[INT0:.*]] = torch.constant.int 0
// CHECK: %[[T1:.*]] = torch_c.from_builtin_tensor %[[T0]] : tensor<2x1x2x1x2xf32> -> !torch.vtensor<[2,1,2,1,2],f32>
// CHECK: return %[[T1]] : !torch.vtensor<[2,1,2,1,2],f32>
func.func @torch.aten.squeeze.dim$0$static(%arg0: !torch.vtensor<[2,1,2,1,2],f32>) -> !torch.vtensor<[2,1,2,1,2],f32> {
%int0 = torch.constant.int 0
%0 = torch.aten.squeeze.dim %arg0, %int0 : !torch.vtensor<[2,1,2,1,2],f32>, !torch.int -> !torch.vtensor<[2,1,2,1,2],f32>
return %0 : !torch.vtensor<[2,1,2,1,2],f32>
}
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// -----
[MHLO] Add [un]squeeze op patterns (#1099) * [MHLO] Add [un]squeeze op patterns * Conform to llvm coding standard * minor update
2022-07-25 23:28:48 +08:00
// CHECK-LABEL: func.func @torch.aten.squeeze.dim$1(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,1,?,1,?],f32>) -> !torch.vtensor<[?,?,1,?],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,1,?,1,?],f32> -> tensor<?x1x?x1x?xf32>
// CHECK: %[[INT1:.*]] = torch.constant.int 1
// CHECK: %[[C0:.*]] = arith.constant 0 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM:.*]] = tensor.dim %[[T0]], %[[C0]] : tensor<?x1x?x1x?xf32>
// CHECK: %[[T1:.*]] = arith.index_cast %[[DIM]] : index to i64
[MHLO] Add [un]squeeze op patterns (#1099) * [MHLO] Add [un]squeeze op patterns * Conform to llvm coding standard * minor update
2022-07-25 23:28:48 +08:00
// CHECK: %[[C2:.*]] = arith.constant 2 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM_0:.*]] = tensor.dim %[[T0]], %[[C2]] : tensor<?x1x?x1x?xf32>
// CHECK: %[[T2:.*]] = arith.index_cast %[[DIM_0]] : index to i64
[MHLO] Add [un]squeeze op patterns (#1099) * [MHLO] Add [un]squeeze op patterns * Conform to llvm coding standard * minor update
2022-07-25 23:28:48 +08:00
// CHECK: %[[C3:.*]] = arith.constant 3 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM_1:.*]] = tensor.dim %[[T0]], %[[C3]] : tensor<?x1x?x1x?xf32>
// CHECK: %[[T3:.*]] = arith.index_cast %[[DIM_1]] : index to i64
[MHLO] Add [un]squeeze op patterns (#1099) * [MHLO] Add [un]squeeze op patterns * Conform to llvm coding standard * minor update
2022-07-25 23:28:48 +08:00
// CHECK: %[[C4:.*]] = arith.constant 4 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM_2:.*]] = tensor.dim %[[T0]], %[[C4]] : tensor<?x1x?x1x?xf32>
// CHECK: %[[T4:.*]] = arith.index_cast %[[DIM_2]] : index to i64
// CHECK: %[[FROM_ELEMENTS:.*]] = tensor.from_elements %[[T1]], %[[T2]], %[[T3]], %[[T4]] : tensor<4xi64>
mhlo: migrate conversion to stablehlo (#1840) This patch replaces all MHLO operations with their StableHLO counterparts and adds a validation pass to ensure that no MHLO operations remain before translating all Stablehlo operations to the MHLO dialect for further lowering to the Linalg dialect. This patch also updates all lit tests so that they refer to the `convert-torch-to-stablehlo` pass and so that they check for StableHLO operations.
2023-02-02 21:29:47 +08:00
// CHECK: %[[T5:.*]] = stablehlo.dynamic_reshape %[[T0]], %[[FROM_ELEMENTS]] : (tensor<?x1x?x1x?xf32>, tensor<4xi64>) -> tensor<?x?x1x?xf32>
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[T6:.*]] = torch_c.from_builtin_tensor %[[T5]] : tensor<?x?x1x?xf32> -> !torch.vtensor<[?,?,1,?],f32>
// CHECK: return %[[T6]] : !torch.vtensor<[?,?,1,?],f32>
[MHLO] Add [un]squeeze op patterns (#1099) * [MHLO] Add [un]squeeze op patterns * Conform to llvm coding standard * minor update
2022-07-25 23:28:48 +08:00
func.func @torch.aten.squeeze.dim$1(%arg0: !torch.vtensor<[?,1,?,1,?],f32>) -> !torch.vtensor<[?,?,1,?],f32> {
%int1 = torch.constant.int 1
%0 = torch.aten.squeeze.dim %arg0, %int1 : !torch.vtensor<[?,1,?,1,?],f32>, !torch.int -> !torch.vtensor<[?,?,1,?],f32>
return %0 : !torch.vtensor<[?,?,1,?],f32>
}
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// -----
[MHLO] Add [un]squeeze op patterns (#1099) * [MHLO] Add [un]squeeze op patterns * Conform to llvm coding standard * minor update
2022-07-25 23:28:48 +08:00
// CHECK-LABEL: func.func @torch.aten.squeeze.dim$from_end(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,1,?,1,?],f32>) -> !torch.vtensor<[?,1,?,?],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,1,?,1,?],f32> -> tensor<?x1x?x1x?xf32>
// CHECK: %[[INT:.*]]-2 = torch.constant.int -2
// CHECK: %[[C0:.*]] = arith.constant 0 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM:.*]] = tensor.dim %[[T0]], %[[C0]] : tensor<?x1x?x1x?xf32>
// CHECK: %[[T1:.*]] = arith.index_cast %[[DIM]] : index to i64
[MHLO] Add [un]squeeze op patterns (#1099) * [MHLO] Add [un]squeeze op patterns * Conform to llvm coding standard * minor update
2022-07-25 23:28:48 +08:00
// CHECK: %[[C1:.*]] = arith.constant 1 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM_0:.*]] = tensor.dim %[[T0]], %[[C1]] : tensor<?x1x?x1x?xf32>
// CHECK: %[[T2:.*]] = arith.index_cast %[[DIM_0]] : index to i64
[MHLO] Add [un]squeeze op patterns (#1099) * [MHLO] Add [un]squeeze op patterns * Conform to llvm coding standard * minor update
2022-07-25 23:28:48 +08:00
// CHECK: %[[C2:.*]] = arith.constant 2 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM_1:.*]] = tensor.dim %[[T0]], %[[C2]] : tensor<?x1x?x1x?xf32>
// CHECK: %[[T3:.*]] = arith.index_cast %[[DIM_1]] : index to i64
[MHLO] Add [un]squeeze op patterns (#1099) * [MHLO] Add [un]squeeze op patterns * Conform to llvm coding standard * minor update
2022-07-25 23:28:48 +08:00
// CHECK: %[[C4:.*]] = arith.constant 4 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM_2:.*]] = tensor.dim %[[T0]], %[[C4]] : tensor<?x1x?x1x?xf32>
// CHECK: %[[T4:.*]] = arith.index_cast %[[DIM_2]] : index to i64
// CHECK: %[[FROM_ELEMENTS:.*]] = tensor.from_elements %[[T1]], %[[T2]], %[[T3]], %[[T4]] : tensor<4xi64>
mhlo: migrate conversion to stablehlo (#1840) This patch replaces all MHLO operations with their StableHLO counterparts and adds a validation pass to ensure that no MHLO operations remain before translating all Stablehlo operations to the MHLO dialect for further lowering to the Linalg dialect. This patch also updates all lit tests so that they refer to the `convert-torch-to-stablehlo` pass and so that they check for StableHLO operations.
2023-02-02 21:29:47 +08:00
// CHECK: %[[T5:.*]] = stablehlo.dynamic_reshape %[[T0]], %[[FROM_ELEMENTS]] : (tensor<?x1x?x1x?xf32>, tensor<4xi64>) -> tensor<?x1x?x?xf32>
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[T6:.*]] = torch_c.from_builtin_tensor %[[T5]] : tensor<?x1x?x?xf32> -> !torch.vtensor<[?,1,?,?],f32>
// CHECK: return %[[T6]] : !torch.vtensor<[?,1,?,?],f32>
[MHLO] Add [un]squeeze op patterns (#1099) * [MHLO] Add [un]squeeze op patterns * Conform to llvm coding standard * minor update
2022-07-25 23:28:48 +08:00
func.func @torch.aten.squeeze.dim$from_end(%arg0: !torch.vtensor<[?,1,?,1,?],f32>) -> !torch.vtensor<[?,1,?,?],f32> {
%int-2 = torch.constant.int -2
%0 = torch.aten.squeeze.dim %arg0, %int-2 : !torch.vtensor<[?,1,?,1,?],f32>, !torch.int -> !torch.vtensor<[?,1,?,?],f32>
return %0 : !torch.vtensor<[?,1,?,?],f32>
}
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// -----
[MHLO] Add [un]squeeze op patterns (#1099) * [MHLO] Add [un]squeeze op patterns * Conform to llvm coding standard * minor update
2022-07-25 23:28:48 +08:00
// CHECK-LABEL: func.func @torch.aten.squeeze$static(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[2,1,2,1,2],f32>) -> !torch.vtensor<[2,2,2],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[2,1,2,1,2],f32> -> tensor<2x1x2x1x2xf32>
// CHECK: %[[C0:.*]] = arith.constant 0 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM:.*]] = tensor.dim %[[T0]], %[[C0]] : tensor<2x1x2x1x2xf32>
// CHECK: %[[T1:.*]] = arith.index_cast %[[DIM]] : index to i64
[MHLO] Add [un]squeeze op patterns (#1099) * [MHLO] Add [un]squeeze op patterns * Conform to llvm coding standard * minor update
2022-07-25 23:28:48 +08:00
// CHECK: %[[C2:.*]] = arith.constant 2 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM_0:.*]] = tensor.dim %[[T0]], %[[C2]] : tensor<2x1x2x1x2xf32>
// CHECK: %[[T2:.*]] = arith.index_cast %[[DIM_0]] : index to i64
[MHLO] Add [un]squeeze op patterns (#1099) * [MHLO] Add [un]squeeze op patterns * Conform to llvm coding standard * minor update
2022-07-25 23:28:48 +08:00
// CHECK: %[[C4:.*]] = arith.constant 4 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM_1:.*]] = tensor.dim %[[T0]], %[[C4]] : tensor<2x1x2x1x2xf32>
// CHECK: %[[T3:.*]] = arith.index_cast %[[DIM_1]] : index to i64
// CHECK: %[[FROM_ELEMENTS:.*]] = tensor.from_elements %[[T1]], %[[T2]], %[[T3]] : tensor<3xi64>
mhlo: migrate conversion to stablehlo (#1840) This patch replaces all MHLO operations with their StableHLO counterparts and adds a validation pass to ensure that no MHLO operations remain before translating all Stablehlo operations to the MHLO dialect for further lowering to the Linalg dialect. This patch also updates all lit tests so that they refer to the `convert-torch-to-stablehlo` pass and so that they check for StableHLO operations.
2023-02-02 21:29:47 +08:00
// CHECK: %[[T4:.*]] = stablehlo.dynamic_reshape %[[T0]], %[[FROM_ELEMENTS]] : (tensor<2x1x2x1x2xf32>, tensor<3xi64>) -> tensor<2x2x2xf32>
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[T5:.*]] = torch_c.from_builtin_tensor %[[T4]] : tensor<2x2x2xf32> -> !torch.vtensor<[2,2,2],f32>
// CHECK: return %[[T5]] : !torch.vtensor<[2,2,2],f32>
[MHLO] Add [un]squeeze op patterns (#1099) * [MHLO] Add [un]squeeze op patterns * Conform to llvm coding standard * minor update
2022-07-25 23:28:48 +08:00
func.func @torch.aten.squeeze$static(%arg0: !torch.vtensor<[2,1,2,1,2],f32>) -> !torch.vtensor<[2,2,2],f32> {
%0 = torch.aten.squeeze %arg0 : !torch.vtensor<[2,1,2,1,2],f32> -> !torch.vtensor<[2,2,2],f32>
return %0 : !torch.vtensor<[2,2,2],f32>
}
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// -----
[MHLO] Add [un]squeeze op patterns (#1099) * [MHLO] Add [un]squeeze op patterns * Conform to llvm coding standard * minor update
2022-07-25 23:28:48 +08:00
// CHECK-LABEL: func.func @torch.aten.unsqueeze$dim$0(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?,?,?],f32>) -> !torch.vtensor<[1,?,?,?,?],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?,?,?],f32> -> tensor<?x?x?x?xf32>
// CHECK: %[[INT0:.*]] = torch.constant.int 0
// CHECK: %[[C0:.*]] = arith.constant 0 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM:.*]] = tensor.dim %[[T0]], %[[C0]] : tensor<?x?x?x?xf32>
// CHECK: %[[T1:.*]] = arith.index_cast %[[DIM]] : index to i64
[MHLO] Add [un]squeeze op patterns (#1099) * [MHLO] Add [un]squeeze op patterns * Conform to llvm coding standard * minor update
2022-07-25 23:28:48 +08:00
// CHECK: %[[C1:.*]] = arith.constant 1 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM_0:.*]] = tensor.dim %[[T0]], %[[C1]] : tensor<?x?x?x?xf32>
// CHECK: %[[T2:.*]] = arith.index_cast %[[DIM_0]] : index to i64
[MHLO] Add [un]squeeze op patterns (#1099) * [MHLO] Add [un]squeeze op patterns * Conform to llvm coding standard * minor update
2022-07-25 23:28:48 +08:00
// CHECK: %[[C2:.*]] = arith.constant 2 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM_1:.*]] = tensor.dim %[[T0]], %[[C2]] : tensor<?x?x?x?xf32>
// CHECK: %[[T3:.*]] = arith.index_cast %[[DIM_1]] : index to i64
[MHLO] Add [un]squeeze op patterns (#1099) * [MHLO] Add [un]squeeze op patterns * Conform to llvm coding standard * minor update
2022-07-25 23:28:48 +08:00
// CHECK: %[[C3:.*]] = arith.constant 3 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM_2:.*]] = tensor.dim %[[T0]], %[[C3]] : tensor<?x?x?x?xf32>
// CHECK: %[[T4:.*]] = arith.index_cast %[[DIM_2]] : index to i64
[MHLO] Add [un]squeeze op patterns (#1099) * [MHLO] Add [un]squeeze op patterns * Conform to llvm coding standard * minor update
2022-07-25 23:28:48 +08:00
// CHECK: %[[C1_I64:.*]] = arith.constant 1 : i64
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[FROM_ELEMENTS:.*]] = tensor.from_elements %[[C1_I64]], %[[T1]], %[[T2]], %[[T3]], %[[T4]] : tensor<5xi64>
mhlo: migrate conversion to stablehlo (#1840) This patch replaces all MHLO operations with their StableHLO counterparts and adds a validation pass to ensure that no MHLO operations remain before translating all Stablehlo operations to the MHLO dialect for further lowering to the Linalg dialect. This patch also updates all lit tests so that they refer to the `convert-torch-to-stablehlo` pass and so that they check for StableHLO operations.
2023-02-02 21:29:47 +08:00
// CHECK: %[[T5:.*]] = stablehlo.dynamic_reshape %[[T0]], %[[FROM_ELEMENTS]] : (tensor<?x?x?x?xf32>, tensor<5xi64>) -> tensor<1x?x?x?x?xf32>
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[T6:.*]] = torch_c.from_builtin_tensor %[[T5]] : tensor<1x?x?x?x?xf32> -> !torch.vtensor<[1,?,?,?,?],f32>
// CHECK: return %[[T6]] : !torch.vtensor<[1,?,?,?,?],f32>
[MHLO] Add [un]squeeze op patterns (#1099) * [MHLO] Add [un]squeeze op patterns * Conform to llvm coding standard * minor update
2022-07-25 23:28:48 +08:00
func.func @torch.aten.unsqueeze$dim$0(%arg0: !torch.vtensor<[?,?,?,?],f32>) -> !torch.vtensor<[1,?,?,?,?],f32> {
%int0 = torch.constant.int 0
%0 = torch.aten.unsqueeze %arg0, %int0 : !torch.vtensor<[?,?,?,?],f32>, !torch.int -> !torch.vtensor<[1,?,?,?,?],f32>
return %0 : !torch.vtensor<[1,?,?,?,?],f32>
}
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// -----
[MHLO] Add [un]squeeze op patterns (#1099) * [MHLO] Add [un]squeeze op patterns * Conform to llvm coding standard * minor update
2022-07-25 23:28:48 +08:00
// CHECK-LABEL: func.func @torch.aten.unsqueeze$dim$1(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?,?,?],f32>) -> !torch.vtensor<[?,1,?,?,?],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?,?,?],f32> -> tensor<?x?x?x?xf32>
// CHECK: %[[INT1:.*]] = torch.constant.int 1
// CHECK: %[[C0:.*]] = arith.constant 0 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM:.*]] = tensor.dim %[[T0]], %[[C0]] : tensor<?x?x?x?xf32>
// CHECK: %[[T1:.*]] = arith.index_cast %[[DIM]] : index to i64
[MHLO] Add [un]squeeze op patterns (#1099) * [MHLO] Add [un]squeeze op patterns * Conform to llvm coding standard * minor update
2022-07-25 23:28:48 +08:00
// CHECK: %[[C1:.*]] = arith.constant 1 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM_0:.*]] = tensor.dim %[[T0]], %[[C1]] : tensor<?x?x?x?xf32>
// CHECK: %[[T2:.*]] = arith.index_cast %[[DIM_0]] : index to i64
[MHLO] Add [un]squeeze op patterns (#1099) * [MHLO] Add [un]squeeze op patterns * Conform to llvm coding standard * minor update
2022-07-25 23:28:48 +08:00
// CHECK: %[[C2:.*]] = arith.constant 2 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM_1:.*]] = tensor.dim %[[T0]], %[[C2]] : tensor<?x?x?x?xf32>
// CHECK: %[[T3:.*]] = arith.index_cast %[[DIM_1]] : index to i64
[MHLO] Add [un]squeeze op patterns (#1099) * [MHLO] Add [un]squeeze op patterns * Conform to llvm coding standard * minor update
2022-07-25 23:28:48 +08:00
// CHECK: %[[C3:.*]] = arith.constant 3 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM_2:.*]] = tensor.dim %[[T0]], %[[C3]] : tensor<?x?x?x?xf32>
// CHECK: %[[T4:.*]] = arith.index_cast %[[DIM_2]] : index to i64
[MHLO] Add [un]squeeze op patterns (#1099) * [MHLO] Add [un]squeeze op patterns * Conform to llvm coding standard * minor update
2022-07-25 23:28:48 +08:00
// CHECK: %[[C1_I64:.*]] = arith.constant 1 : i64
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[FROM_ELEMENTS:.*]] = tensor.from_elements %[[T1]], %[[C1_I64]], %[[T2]], %[[T3]], %[[T4]] : tensor<5xi64>
mhlo: migrate conversion to stablehlo (#1840) This patch replaces all MHLO operations with their StableHLO counterparts and adds a validation pass to ensure that no MHLO operations remain before translating all Stablehlo operations to the MHLO dialect for further lowering to the Linalg dialect. This patch also updates all lit tests so that they refer to the `convert-torch-to-stablehlo` pass and so that they check for StableHLO operations.
2023-02-02 21:29:47 +08:00
// CHECK: %[[T5:.*]] = stablehlo.dynamic_reshape %[[T0]], %[[FROM_ELEMENTS]] : (tensor<?x?x?x?xf32>, tensor<5xi64>) -> tensor<?x1x?x?x?xf32>
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[T6:.*]] = torch_c.from_builtin_tensor %[[T5]] : tensor<?x1x?x?x?xf32> -> !torch.vtensor<[?,1,?,?,?],f32>
// CHECK: return %[[T6]] : !torch.vtensor<[?,1,?,?,?],f32>
[MHLO] Add [un]squeeze op patterns (#1099) * [MHLO] Add [un]squeeze op patterns * Conform to llvm coding standard * minor update
2022-07-25 23:28:48 +08:00
func.func @torch.aten.unsqueeze$dim$1(%arg0: !torch.vtensor<[?,?,?,?],f32>) -> !torch.vtensor<[?,1,?,?,?],f32> {
%int1 = torch.constant.int 1
%0 = torch.aten.unsqueeze %arg0, %int1 : !torch.vtensor<[?,?,?,?],f32>, !torch.int -> !torch.vtensor<[?,1,?,?,?],f32>
return %0 : !torch.vtensor<[?,1,?,?,?],f32>
}
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// -----
[MHLO] Add [un]squeeze op patterns (#1099) * [MHLO] Add [un]squeeze op patterns * Conform to llvm coding standard * minor update
2022-07-25 23:28:48 +08:00
// CHECK-LABEL: func.func @torch.aten.unsqueeze$from_end(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?,?,?],f32>) -> !torch.vtensor<[?,?,?,1,?],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?,?,?],f32> -> tensor<?x?x?x?xf32>
// CHECK: %[[INT:.*]]-2 = torch.constant.int -2
// CHECK: %[[C0:.*]] = arith.constant 0 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM:.*]] = tensor.dim %[[T0]], %[[C0]] : tensor<?x?x?x?xf32>
// CHECK: %[[T1:.*]] = arith.index_cast %[[DIM]] : index to i64
[MHLO] Add [un]squeeze op patterns (#1099) * [MHLO] Add [un]squeeze op patterns * Conform to llvm coding standard * minor update
2022-07-25 23:28:48 +08:00
// CHECK: %[[C1:.*]] = arith.constant 1 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM_0:.*]] = tensor.dim %[[T0]], %[[C1]] : tensor<?x?x?x?xf32>
// CHECK: %[[T2:.*]] = arith.index_cast %[[DIM_0]] : index to i64
[MHLO] Add [un]squeeze op patterns (#1099) * [MHLO] Add [un]squeeze op patterns * Conform to llvm coding standard * minor update
2022-07-25 23:28:48 +08:00
// CHECK: %[[C2:.*]] = arith.constant 2 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM_1:.*]] = tensor.dim %[[T0]], %[[C2]] : tensor<?x?x?x?xf32>
// CHECK: %[[T3:.*]] = arith.index_cast %[[DIM_1]] : index to i64
[MHLO] Add [un]squeeze op patterns (#1099) * [MHLO] Add [un]squeeze op patterns * Conform to llvm coding standard * minor update
2022-07-25 23:28:48 +08:00
// CHECK: %[[C3:.*]] = arith.constant 3 : index
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[DIM_2:.*]] = tensor.dim %[[T0]], %[[C3]] : tensor<?x?x?x?xf32>
// CHECK: %[[T4:.*]] = arith.index_cast %[[DIM_2]] : index to i64
[MHLO] Add [un]squeeze op patterns (#1099) * [MHLO] Add [un]squeeze op patterns * Conform to llvm coding standard * minor update
2022-07-25 23:28:48 +08:00
// CHECK: %[[C1_I64:.*]] = arith.constant 1 : i64
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[FROM_ELEMENTS:.*]] = tensor.from_elements %[[T1]], %[[T2]], %[[T3]], %[[C1_I64]], %[[T4]] : tensor<5xi64>
mhlo: migrate conversion to stablehlo (#1840) This patch replaces all MHLO operations with their StableHLO counterparts and adds a validation pass to ensure that no MHLO operations remain before translating all Stablehlo operations to the MHLO dialect for further lowering to the Linalg dialect. This patch also updates all lit tests so that they refer to the `convert-torch-to-stablehlo` pass and so that they check for StableHLO operations.
2023-02-02 21:29:47 +08:00
// CHECK: %[[T5:.*]] = stablehlo.dynamic_reshape %[[T0]], %[[FROM_ELEMENTS]] : (tensor<?x?x?x?xf32>, tensor<5xi64>) -> tensor<?x?x?x1x?xf32>
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00
// CHECK: %[[T6:.*]] = torch_c.from_builtin_tensor %[[T5]] : tensor<?x?x?x1x?xf32> -> !torch.vtensor<[?,?,?,1,?],f32>
// CHECK: return %[[T6]] : !torch.vtensor<[?,?,?,1,?],f32>
[MHLO] Add [un]squeeze op patterns (#1099) * [MHLO] Add [un]squeeze op patterns * Conform to llvm coding standard * minor update
2022-07-25 23:28:48 +08:00
func.func @torch.aten.unsqueeze$from_end(%arg0: !torch.vtensor<[?,?,?,?],f32>) -> !torch.vtensor<[?,?,?,1,?],f32> {
%int-2 = torch.constant.int -2
%0 = torch.aten.unsqueeze %arg0, %int-2 : !torch.vtensor<[?,?,?,?],f32>, !torch.int -> !torch.vtensor<[?,?,?,1,?],f32>
return %0 : !torch.vtensor<[?,?,?,1,?],f32>
}
build: update llvm tag to 74fb770d (#1539) * build: update llvm tag to 74fb770d This commit makes the following changes needed to update bump LLVM: + replace usages of `tensor::createPadScalarOp`, see https://reviews.llvm.org/D136493 + Update file checks
2022-11-01 15:27:09 +08:00