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/elementwise.mlir

627 lines
43 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] add aten::gelu op pattern (#1127) add aten::gelu op pattern, and moved some unit tests from basic.mlir to elementwise.mlir
2022-08-02 15:01:30 +08:00
// CHECK-LABEL: func.func @torch.aten.gelu(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?],f32> -> tensor<?x?xf32>
[MHLO] support non-constant torch scalar in BasicOps (#1134) 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
2022-08-03 08:16:31 +08:00
// CHECK: %[[STR:.*]] = torch.constant.str "none"
[MHLO] add aten::gelu op pattern (#1127) add aten::gelu op pattern, and moved some unit tests from basic.mlir to elementwise.mlir
2022-08-02 15:01:30 +08:00
// CHECK: %[[T1:.*]] = "chlo.constant_like"(%[[T0]]) {value = 1.000000e+00 : f32} : (tensor<?x?xf32>) -> tensor<?x?xf32>
// CHECK: %[[T2:.*]] = "chlo.constant_like"(%[[T0]]) {value = 2.000000e+00 : f32} : (tensor<?x?xf32>) -> tensor<?x?xf32>
// CHECK: %[[T3:.*]] = "chlo.constant_like"(%[[T0]]) {value = 5.000000e-01 : f32} : (tensor<?x?xf32>) -> tensor<?x?xf32>
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.rsqrt %[[T2]] : tensor<?x?xf32>
// CHECK: %[[T5:.*]] = stablehlo.multiply %[[T0]], %[[T4]] : tensor<?x?xf32>
[MHLO] add aten::gelu op pattern (#1127) add aten::gelu op pattern, and moved some unit tests from basic.mlir to elementwise.mlir
2022-08-02 15:01:30 +08:00
// CHECK: %[[T6:.*]] = chlo.erf %[[T5]] : tensor<?x?xf32> -> tensor<?x?xf32>
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.add %[[T6]], %[[T1]] : tensor<?x?xf32>
// CHECK: %[[T8:.*]] = stablehlo.multiply %[[T7]], %[[T3]] : tensor<?x?xf32>
// CHECK: %[[T9:.*]] = stablehlo.multiply %[[T0]], %[[T8]] : tensor<?x?xf32>
[MHLO] add aten::gelu op pattern (#1127) add aten::gelu op pattern, and moved some unit tests from basic.mlir to elementwise.mlir
2022-08-02 15:01:30 +08:00
// CHECK: %[[T10:.*]] = torch_c.from_builtin_tensor %[[T9]] : tensor<?x?xf32> -> !torch.vtensor<[?,?],f32>
// CHECK: return %[[T10]] : !torch.vtensor<[?,?],f32>
func.func @torch.aten.gelu(%arg0: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
%str = torch.constant.str "none"
%0 = torch.aten.gelu %arg0, %str : !torch.vtensor<[?,?],f32>, !torch.str -> !torch.vtensor<[?,?],f32>
return %0 : !torch.vtensor<[?,?],f32>
}
[MHLO] support non-constant torch scalar in BasicOps (#1134) 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
2022-08-03 08:16:31 +08:00
// -----
[MHLO] add aten::gelu op pattern (#1127) add aten::gelu op pattern, and moved some unit tests from basic.mlir to elementwise.mlir
2022-08-02 15:01:30 +08:00
[MHLO] support non-constant torch scalar in BasicOps (#1134) 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
2022-08-03 08:16:31 +08:00
// CHECK-LABEL: func.func @torch.aten.tanh$basic(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?],f32> -> tensor<?x?xf32>
mhlo: migrate conversion to stablehlo (#1840) This patch replaces all MHLO operations with their StableHLO counterparts and adds a validation pass to ensure that no MHLO operations remain before translating all Stablehlo operations to the MHLO dialect for further lowering to the Linalg dialect. This patch also updates all lit tests so that they refer to the `convert-torch-to-stablehlo` pass and so that they check for StableHLO operations.
2023-02-02 21:29:47 +08:00
// CHECK: %[[T1:.*]] = stablehlo.tanh %[[T0]] : tensor<?x?xf32>
[MHLO] support non-constant torch scalar in BasicOps (#1134) See RFC https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-08-03 08:16:31 +08:00
// CHECK: %[[T2:.*]] = torch_c.from_builtin_tensor %[[T1]] : tensor<?x?xf32> -> !torch.vtensor<[?,?],f32>
// CHECK: return %[[T2]] : !torch.vtensor<[?,?],f32>
[MHLO] add aten::gelu op pattern (#1127) add aten::gelu op pattern, and moved some unit tests from basic.mlir to elementwise.mlir
2022-08-02 15:01:30 +08:00
func.func @torch.aten.tanh$basic(%arg0: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
%0 = torch.aten.tanh %arg0 : !torch.vtensor<[?,?],f32> -> !torch.vtensor<[?,?],f32>
return %0 : !torch.vtensor<[?,?],f32>
}
// -----
[MHLO] support non-constant torch scalar in BasicOps (#1134) 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
2022-08-03 08:16:31 +08:00
// CHECK-LABEL: func.func @torch.aten.log$basic(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?],f32> -> tensor<?x?xf32>
mhlo: migrate conversion to stablehlo (#1840) This patch replaces all MHLO operations with their StableHLO counterparts and adds a validation pass to ensure that no MHLO operations remain before translating all Stablehlo operations to the MHLO dialect for further lowering to the Linalg dialect. This patch also updates all lit tests so that they refer to the `convert-torch-to-stablehlo` pass and so that they check for StableHLO operations.
2023-02-02 21:29:47 +08:00
// CHECK: %[[T1:.*]] = stablehlo.log %[[T0]] : tensor<?x?xf32>
[MHLO] support non-constant torch scalar in BasicOps (#1134) See RFC https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-08-03 08:16:31 +08:00
// CHECK: %[[T2:.*]] = torch_c.from_builtin_tensor %[[T1]] : tensor<?x?xf32> -> !torch.vtensor<[?,?],f32>
// CHECK: return %[[T2]] : !torch.vtensor<[?,?],f32>
[MHLO] add aten::gelu op pattern (#1127) add aten::gelu op pattern, and moved some unit tests from basic.mlir to elementwise.mlir
2022-08-02 15:01:30 +08:00
func.func @torch.aten.log$basic(%arg0: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
%0 = torch.aten.log %arg0 : !torch.vtensor<[?,?],f32> -> !torch.vtensor<[?,?],f32>
return %0 : !torch.vtensor<[?,?],f32>
}
// -----
[MHLO] support non-constant torch scalar in BasicOps (#1134) 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
2022-08-03 08:16:31 +08:00
// CHECK-LABEL: func.func @torch.aten.exp$basic(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?],f32> -> tensor<?x?xf32>
mhlo: migrate conversion to stablehlo (#1840) This patch replaces all MHLO operations with their StableHLO counterparts and adds a validation pass to ensure that no MHLO operations remain before translating all Stablehlo operations to the MHLO dialect for further lowering to the Linalg dialect. This patch also updates all lit tests so that they refer to the `convert-torch-to-stablehlo` pass and so that they check for StableHLO operations.
2023-02-02 21:29:47 +08:00
// CHECK: %[[T1:.*]] = stablehlo.exponential %[[T0]] : tensor<?x?xf32>
[MHLO] support non-constant torch scalar in BasicOps (#1134) See RFC https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-08-03 08:16:31 +08:00
// CHECK: %[[T2:.*]] = torch_c.from_builtin_tensor %[[T1]] : tensor<?x?xf32> -> !torch.vtensor<[?,?],f32>
// CHECK: return %[[T2]] : !torch.vtensor<[?,?],f32>
[MHLO] add aten::gelu op pattern (#1127) add aten::gelu op pattern, and moved some unit tests from basic.mlir to elementwise.mlir
2022-08-02 15:01:30 +08:00
func.func @torch.aten.exp$basic(%arg0: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
%0 = torch.aten.exp %arg0 : !torch.vtensor<[?,?],f32> -> !torch.vtensor<[?,?],f32>
return %0 : !torch.vtensor<[?,?],f32>
}
// -----
[MHLO] support non-constant torch scalar in BasicOps (#1134) 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
2022-08-03 08:16:31 +08:00
// CHECK-LABEL: func.func @torch.aten.neg$basic(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?],f32> -> tensor<?x?xf32>
mhlo: migrate conversion to stablehlo (#1840) This patch replaces all MHLO operations with their StableHLO counterparts and adds a validation pass to ensure that no MHLO operations remain before translating all Stablehlo operations to the MHLO dialect for further lowering to the Linalg dialect. This patch also updates all lit tests so that they refer to the `convert-torch-to-stablehlo` pass and so that they check for StableHLO operations.
2023-02-02 21:29:47 +08:00
// CHECK: %[[T1:.*]] = stablehlo.negate %[[T0]] : tensor<?x?xf32>
[MHLO] support non-constant torch scalar in BasicOps (#1134) See RFC https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-08-03 08:16:31 +08:00
// CHECK: %[[T2:.*]] = torch_c.from_builtin_tensor %[[T1]] : tensor<?x?xf32> -> !torch.vtensor<[?,?],f32>
// CHECK: return %[[T2]] : !torch.vtensor<[?,?],f32>
[MHLO] add aten::gelu op pattern (#1127) add aten::gelu op pattern, and moved some unit tests from basic.mlir to elementwise.mlir
2022-08-02 15:01:30 +08:00
func.func @torch.aten.neg$basic(%arg0: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
%0 = torch.aten.neg %arg0 : !torch.vtensor<[?,?],f32> -> !torch.vtensor<[?,?],f32>
return %0 : !torch.vtensor<[?,?],f32>
}
// -----
[Torch][MHLO] Decompose aten.copy op. Lower aten.rsqrt & sigmoid to mhlo. (#1734)
2022-12-22 10:13:59 +08:00
// CHECK-LABEL: func.func @torch.aten.rsqrt$basic(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?],f32> -> tensor<?x?xf32>
mhlo: migrate conversion to stablehlo (#1840) This patch replaces all MHLO operations with their StableHLO counterparts and adds a validation pass to ensure that no MHLO operations remain before translating all Stablehlo operations to the MHLO dialect for further lowering to the Linalg dialect. This patch also updates all lit tests so that they refer to the `convert-torch-to-stablehlo` pass and so that they check for StableHLO operations.
2023-02-02 21:29:47 +08:00
// CHECK: %[[T1:.*]] = stablehlo.rsqrt %[[T0]] : tensor<?x?xf32>
[Torch][MHLO] Decompose aten.copy op. Lower aten.rsqrt & sigmoid to mhlo. (#1734)
2022-12-22 10:13:59 +08:00
// CHECK: %[[T2:.*]] = torch_c.from_builtin_tensor %[[T1]] : tensor<?x?xf32> -> !torch.vtensor<[?,?],f32>
// CHECK: return %[[T2]] : !torch.vtensor<[?,?],f32>
func.func @torch.aten.rsqrt$basic(%arg0: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
%0 = torch.aten.rsqrt %arg0 : !torch.vtensor<[?,?],f32> -> !torch.vtensor<[?,?],f32>
return %0 : !torch.vtensor<[?,?],f32>
}
// -----
// CHECK-LABEL: func.func @torch.aten.sigmoid$basic(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?],f32> -> tensor<?x?xf32>
mhlo: migrate conversion to stablehlo (#1840) This patch replaces all MHLO operations with their StableHLO counterparts and adds a validation pass to ensure that no MHLO operations remain before translating all Stablehlo operations to the MHLO dialect for further lowering to the Linalg dialect. This patch also updates all lit tests so that they refer to the `convert-torch-to-stablehlo` pass and so that they check for StableHLO operations.
2023-02-02 21:29:47 +08:00
// CHECK: %[[T1:.*]] = stablehlo.logistic %[[T0]] : tensor<?x?xf32>
[Torch][MHLO] Decompose aten.copy op. Lower aten.rsqrt & sigmoid to mhlo. (#1734)
2022-12-22 10:13:59 +08:00
// CHECK: %[[T2:.*]] = torch_c.from_builtin_tensor %[[T1]] : tensor<?x?xf32> -> !torch.vtensor<[?,?],f32>
// CHECK: return %[[T2]] : !torch.vtensor<[?,?],f32>
func.func @torch.aten.sigmoid$basic(%arg0: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
%0 = torch.aten.sigmoid %arg0 : !torch.vtensor<[?,?],f32> -> !torch.vtensor<[?,?],f32>
return %0 : !torch.vtensor<[?,?],f32>
}
// -----
[MHLO] support non-constant torch scalar in BasicOps (#1134) 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
2022-08-03 08:16:31 +08:00
// CHECK-LABEL: func.func @torch.aten.addscalar$basic(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?],f32> -> tensor<?x?xf32>
// CHECK: %[[INT9:.*]] = torch.constant.int 9
// CHECK: %[[T1:.*]] = torch_c.to_i64 %[[INT9]]
// CHECK: %[[INT1:.*]] = torch.constant.int 1
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]] : tensor<1xi64>
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.convert %[[FROM_ELEMENTS]] : (tensor<1xi64>) -> tensor<1xf32>
// CHECK: %[[T3:.*]] = stablehlo.reshape %[[T2]] : (tensor<1xf32>) -> tensor<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
// CHECK: %[[T4:.*]] = chlo.broadcast_add %[[T0]], %[[T3]] : (tensor<?x?xf32>, tensor<f32>) -> tensor<?x?xf32>
// CHECK: %[[T5:.*]] = torch_c.from_builtin_tensor %[[T4]] : tensor<?x?xf32> -> !torch.vtensor<[?,?],f32>
// CHECK: return %[[T5]] : !torch.vtensor<[?,?],f32>
[MHLO] add aten::gelu op pattern (#1127) add aten::gelu op pattern, and moved some unit tests from basic.mlir to elementwise.mlir
2022-08-02 15:01:30 +08:00
func.func @torch.aten.addscalar$basic(%arg0: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
%int9 = torch.constant.int 9
%int1 = torch.constant.int 1
%0 = torch.aten.add.Scalar %arg0, %int9, %int1 : !torch.vtensor<[?,?],f32>, !torch.int, !torch.int -> !torch.vtensor<[?,?],f32>
return %0 : !torch.vtensor<[?,?],f32>
}
// -----
[MHLO] support non-constant torch scalar in BasicOps (#1134) 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
2022-08-03 08:16:31 +08:00
// CHECK-LABEL: func.func @torch.aten.addscalar$alpha(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?],f32> -> tensor<?x?xf32>
// CHECK: %[[INT9:.*]] = torch.constant.int 9
// CHECK: %[[T1:.*]] = torch_c.to_i64 %[[INT9]]
// CHECK: %[[INT2:.*]] = torch.constant.int 2
// CHECK: %[[T2:.*]] = torch_c.to_i64 %[[INT2]]
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]] : tensor<1xi64>
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: %[[T3:.*]] = stablehlo.convert %[[FROM_ELEMENTS]] : (tensor<1xi64>) -> tensor<1xf32>
// CHECK: %[[T4:.*]] = stablehlo.reshape %[[T3]] : (tensor<1xf32>) -> tensor<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
// CHECK: %[[FROM_ELEMENTS_0:.*]] = tensor.from_elements %[[T2]] : tensor<1xi64>
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.convert %[[FROM_ELEMENTS_0]] : (tensor<1xi64>) -> tensor<1xf32>
// CHECK: %[[T6:.*]] = stablehlo.reshape %[[T5]] : (tensor<1xf32>) -> tensor<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
// CHECK: %[[T7:.*]] = chlo.broadcast_multiply %[[T4]], %[[T6]] : (tensor<f32>, tensor<f32>) -> tensor<f32>
// CHECK: %[[T8:.*]] = chlo.broadcast_add %[[T0]], %[[T7]] : (tensor<?x?xf32>, tensor<f32>) -> tensor<?x?xf32>
// CHECK: %[[T9:.*]] = torch_c.from_builtin_tensor %[[T8]] : tensor<?x?xf32> -> !torch.vtensor<[?,?],f32>
// CHECK: return %[[T9]] : !torch.vtensor<[?,?],f32>
[MHLO] add aten::gelu op pattern (#1127) add aten::gelu op pattern, and moved some unit tests from basic.mlir to elementwise.mlir
2022-08-02 15:01:30 +08:00
func.func @torch.aten.addscalar$alpha(%arg0: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
%int9 = torch.constant.int 9
%int2 = torch.constant.int 2
%0 = torch.aten.add.Scalar %arg0, %int9, %int2 : !torch.vtensor<[?,?],f32>, !torch.int, !torch.int -> !torch.vtensor<[?,?],f32>
return %0 : !torch.vtensor<[?,?],f32>
}
// -----
[MHLO] support non-constant torch scalar in BasicOps (#1134) 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
2022-08-03 08:16:31 +08:00
// CHECK-LABEL: func.func @torch.aten.addtensor$basic(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?],f32>, %[[ARG1:.*]]: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?],f32> -> tensor<?x?xf32>
// CHECK: %[[T1:.*]] = torch_c.to_builtin_tensor %[[ARG1]] : !torch.vtensor<[?,?],f32> -> tensor<?x?xf32>
// CHECK: %[[INT1:.*]] = torch.constant.int 1
// CHECK: %[[T2:.*]] = chlo.broadcast_add %[[T0]], %[[T1]] : (tensor<?x?xf32>, tensor<?x?xf32>) -> tensor<?x?xf32>
// CHECK: %[[T3:.*]] = torch_c.from_builtin_tensor %[[T2]] : tensor<?x?xf32> -> !torch.vtensor<[?,?],f32>
// CHECK: return %[[T3]] : !torch.vtensor<[?,?],f32>
[MHLO] add aten::gelu op pattern (#1127) add aten::gelu op pattern, and moved some unit tests from basic.mlir to elementwise.mlir
2022-08-02 15:01:30 +08:00
func.func @torch.aten.addtensor$basic(%arg0: !torch.vtensor<[?,?],f32>, %arg1: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
%int1 = torch.constant.int 1
%0 = torch.aten.add.Tensor %arg0, %arg1, %int1 : !torch.vtensor<[?,?],f32>, !torch.vtensor<[?,?],f32>, !torch.int -> !torch.vtensor<[?,?],f32>
return %0 : !torch.vtensor<[?,?],f32>
}
// -----
[MHLO] support non-constant torch scalar in BasicOps (#1134) 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
2022-08-03 08:16:31 +08:00
// CHECK-LABEL: func.func @torch.aten.addtensor$alpha(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?],f32>, %[[ARG1:.*]]: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?],f32> -> tensor<?x?xf32>
// CHECK: %[[T1:.*]] = torch_c.to_builtin_tensor %[[ARG1]] : !torch.vtensor<[?,?],f32> -> tensor<?x?xf32>
// CHECK: %[[INT2:.*]] = torch.constant.int 2
// CHECK: %[[T2:.*]] = torch_c.to_i64 %[[INT2]]
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]] : tensor<1xi64>
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: %[[T3:.*]] = stablehlo.convert %[[FROM_ELEMENTS]] : (tensor<1xi64>) -> tensor<1xf32>
// CHECK: %[[T4:.*]] = stablehlo.reshape %[[T3]] : (tensor<1xf32>) -> tensor<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
// CHECK: %[[T5:.*]] = chlo.broadcast_multiply %[[T1]], %[[T4]] : (tensor<?x?xf32>, tensor<f32>) -> tensor<?x?xf32>
// CHECK: %[[T6:.*]] = chlo.broadcast_add %[[T0]], %[[T5]] : (tensor<?x?xf32>, tensor<?x?xf32>) -> tensor<?x?xf32>
// CHECK: %[[T7:.*]] = torch_c.from_builtin_tensor %[[T6]] : tensor<?x?xf32> -> !torch.vtensor<[?,?],f32>
// CHECK: return %[[T7]] : !torch.vtensor<[?,?],f32>
[MHLO] add aten::gelu op pattern (#1127) add aten::gelu op pattern, and moved some unit tests from basic.mlir to elementwise.mlir
2022-08-02 15:01:30 +08:00
func.func @torch.aten.addtensor$alpha(%arg0: !torch.vtensor<[?,?],f32>, %arg1: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
%int2 = torch.constant.int 2
%0 = torch.aten.add.Tensor %arg0, %arg1, %int2 : !torch.vtensor<[?,?],f32>, !torch.vtensor<[?,?],f32>, !torch.int -> !torch.vtensor<[?,?],f32>
return %0 : !torch.vtensor<[?,?],f32>
}
// -----
[MHLO] support non-constant torch scalar in BasicOps (#1134) 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
2022-08-03 08:16:31 +08:00
// CHECK-LABEL: func.func @torch.aten.addtensor$promote(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?],si32>, %[[ARG1:.*]]: !torch.vtensor<[?,?],si64>) -> !torch.vtensor<[?,?],si64> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?],si32> -> tensor<?x?xi32>
// CHECK: %[[T1:.*]] = torch_c.to_builtin_tensor %[[ARG1]] : !torch.vtensor<[?,?],si64> -> tensor<?x?xi64>
// CHECK: %[[INT1:.*]] = torch.constant.int 1
mhlo: migrate conversion to stablehlo (#1840) This patch replaces all MHLO operations with their StableHLO counterparts and adds a validation pass to ensure that no MHLO operations remain before translating all Stablehlo operations to the MHLO dialect for further lowering to the Linalg dialect. This patch also updates all lit tests so that they refer to the `convert-torch-to-stablehlo` pass and so that they check for StableHLO operations.
2023-02-02 21:29:47 +08:00
// CHECK: %[[T2:.*]] = stablehlo.convert %[[T0]] : (tensor<?x?xi32>) -> tensor<?x?xi64>
[MHLO] support non-constant torch scalar in BasicOps (#1134) See RFC https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-08-03 08:16:31 +08:00
// CHECK: %[[T3:.*]] = chlo.broadcast_add %[[T2]], %[[T1]] : (tensor<?x?xi64>, tensor<?x?xi64>) -> tensor<?x?xi64>
// CHECK: %[[T4:.*]] = torch_c.from_builtin_tensor %[[T3]] : tensor<?x?xi64> -> !torch.vtensor<[?,?],si64>
// CHECK: return %[[T4]] : !torch.vtensor<[?,?],si64>
[MHLO] add aten::gelu op pattern (#1127) add aten::gelu op pattern, and moved some unit tests from basic.mlir to elementwise.mlir
2022-08-02 15:01:30 +08:00
func.func @torch.aten.addtensor$promote(%arg0: !torch.vtensor<[?,?],si32>, %arg1: !torch.vtensor<[?,?],si64>) -> !torch.vtensor<[?,?],si64> {
%int1 = torch.constant.int 1
%0 = torch.aten.add.Tensor %arg0, %arg1, %int1 : !torch.vtensor<[?,?],si32>, !torch.vtensor<[?,?],si64>, !torch.int -> !torch.vtensor<[?,?],si64>
return %0 : !torch.vtensor<[?,?],si64>
}
// -----
[MHLO] support non-constant torch scalar in BasicOps (#1134) 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
2022-08-03 08:16:31 +08:00
// CHECK-LABEL: func.func @torch.aten.subscalar$basic(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?],f32> -> tensor<?x?xf32>
// CHECK: %[[INT9:.*]] = torch.constant.int 9
// CHECK: %[[T1:.*]] = torch_c.to_i64 %[[INT9]]
// CHECK: %[[INT1:.*]] = torch.constant.int 1
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]] : tensor<1xi64>
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.convert %[[FROM_ELEMENTS]] : (tensor<1xi64>) -> tensor<1xf32>
// CHECK: %[[T3:.*]] = stablehlo.reshape %[[T2]] : (tensor<1xf32>) -> tensor<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
// CHECK: %[[T4:.*]] = chlo.broadcast_subtract %[[T0]], %[[T3]] : (tensor<?x?xf32>, tensor<f32>) -> tensor<?x?xf32>
// CHECK: %[[T5:.*]] = torch_c.from_builtin_tensor %[[T4]] : tensor<?x?xf32> -> !torch.vtensor<[?,?],f32>
// CHECK: return %[[T5]] : !torch.vtensor<[?,?],f32>
[MHLO] add aten::gelu op pattern (#1127) add aten::gelu op pattern, and moved some unit tests from basic.mlir to elementwise.mlir
2022-08-02 15:01:30 +08:00
func.func @torch.aten.subscalar$basic(%arg0: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
%int9 = torch.constant.int 9
%int1 = torch.constant.int 1
%0 = torch.aten.sub.Scalar %arg0, %int9, %int1 : !torch.vtensor<[?,?],f32>, !torch.int, !torch.int -> !torch.vtensor<[?,?],f32>
return %0 : !torch.vtensor<[?,?],f32>
}
// -----
[MHLO] Add AtenRSubScalarOp conversion pattern to MHLO (#1233) * [MHLO] Add AtenRSubScalarOp conversion pattern 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>
2022-08-17 09:07:36 +08:00
// CHECK-LABEL: func.func @torch.aten.rsubscalar$basic(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?],f32> -> tensor<?x?xf32>
// CHECK: %[[INT9:.*]] = torch.constant.int 9
// CHECK: %[[T1:.*]] = torch_c.to_i64 %[[INT9]]
// CHECK: %[[INT1:.*]] = torch.constant.int 1
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]] : tensor<1xi64>
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.convert %[[FROM_ELEMENTS]] : (tensor<1xi64>) -> tensor<1xf32>
// CHECK: %[[T3:.*]] = stablehlo.reshape %[[T2]] : (tensor<1xf32>) -> tensor<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
// CHECK: %[[T4:.*]] = chlo.broadcast_subtract %[[T3]], %[[T0]] : (tensor<f32>, tensor<?x?xf32>) -> tensor<?x?xf32>
// CHECK: %[[T5:.*]] = torch_c.from_builtin_tensor %[[T4]] : tensor<?x?xf32> -> !torch.vtensor<[?,?],f32>
// CHECK: return %[[T5]] : !torch.vtensor<[?,?],f32>
[MHLO] Add AtenRSubScalarOp conversion pattern to MHLO (#1233) * [MHLO] Add AtenRSubScalarOp conversion pattern 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>
2022-08-17 09:07:36 +08:00
func.func @torch.aten.rsubscalar$basic(%arg0: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
%int9 = torch.constant.int 9
%int1 = torch.constant.int 1
%0 = torch.aten.rsub.Scalar %arg0, %int9, %int1 : !torch.vtensor<[?,?],f32>, !torch.int, !torch.int -> !torch.vtensor<[?,?],f32>
return %0 : !torch.vtensor<[?,?],f32>
}
// -----
[MHLO] support non-constant torch scalar in BasicOps (#1134) 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
2022-08-03 08:16:31 +08:00
// CHECK-LABEL: func.func @torch.aten.subscalar$alpha(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?],f32> -> tensor<?x?xf32>
// CHECK: %[[INT9:.*]] = torch.constant.int 9
// CHECK: %[[T1:.*]] = torch_c.to_i64 %[[INT9]]
// CHECK: %[[INT2:.*]] = torch.constant.int 2
// CHECK: %[[T2:.*]] = torch_c.to_i64 %[[INT2]]
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]] : tensor<1xi64>
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: %[[T3:.*]] = stablehlo.convert %[[FROM_ELEMENTS]] : (tensor<1xi64>) -> tensor<1xf32>
// CHECK: %[[T4:.*]] = stablehlo.reshape %[[T3]] : (tensor<1xf32>) -> tensor<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
// CHECK: %[[FROM_ELEMENTS_0:.*]] = tensor.from_elements %[[T2]] : tensor<1xi64>
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.convert %[[FROM_ELEMENTS_0]] : (tensor<1xi64>) -> tensor<1xf32>
// CHECK: %[[T6:.*]] = stablehlo.reshape %[[T5]] : (tensor<1xf32>) -> tensor<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
// CHECK: %[[T7:.*]] = chlo.broadcast_multiply %[[T4]], %[[T6]] : (tensor<f32>, tensor<f32>) -> tensor<f32>
// CHECK: %[[T8:.*]] = chlo.broadcast_subtract %[[T0]], %[[T7]] : (tensor<?x?xf32>, tensor<f32>) -> tensor<?x?xf32>
// CHECK: %[[T9:.*]] = torch_c.from_builtin_tensor %[[T8]] : tensor<?x?xf32> -> !torch.vtensor<[?,?],f32>
// CHECK: return %[[T9]] : !torch.vtensor<[?,?],f32>
[MHLO] add aten::gelu op pattern (#1127) add aten::gelu op pattern, and moved some unit tests from basic.mlir to elementwise.mlir
2022-08-02 15:01:30 +08:00
func.func @torch.aten.subscalar$alpha(%arg0: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
%int9 = torch.constant.int 9
%int2 = torch.constant.int 2
%0 = torch.aten.sub.Scalar %arg0, %int9, %int2 : !torch.vtensor<[?,?],f32>, !torch.int, !torch.int -> !torch.vtensor<[?,?],f32>
return %0 : !torch.vtensor<[?,?],f32>
}
// -----
[MHLO] support non-constant torch scalar in BasicOps (#1134) 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
2022-08-03 08:16:31 +08:00
// CHECK-LABEL: func.func @torch.aten.subtensor$basic(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?],f32>, %[[ARG1:.*]]: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?],f32> -> tensor<?x?xf32>
// CHECK: %[[T1:.*]] = torch_c.to_builtin_tensor %[[ARG1]] : !torch.vtensor<[?,?],f32> -> tensor<?x?xf32>
// CHECK: %[[INT1:.*]] = torch.constant.int 1
// CHECK: %[[T2:.*]] = chlo.broadcast_subtract %[[T0]], %[[T1]] : (tensor<?x?xf32>, tensor<?x?xf32>) -> tensor<?x?xf32>
// CHECK: %[[T3:.*]] = torch_c.from_builtin_tensor %[[T2]] : tensor<?x?xf32> -> !torch.vtensor<[?,?],f32>
// CHECK: return %[[T3]] : !torch.vtensor<[?,?],f32>
[MHLO] add aten::gelu op pattern (#1127) add aten::gelu op pattern, and moved some unit tests from basic.mlir to elementwise.mlir
2022-08-02 15:01:30 +08:00
func.func @torch.aten.subtensor$basic(%arg0: !torch.vtensor<[?,?],f32>, %arg1: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
%int1 = torch.constant.int 1
%0 = torch.aten.sub.Tensor %arg0, %arg1, %int1 : !torch.vtensor<[?,?],f32>, !torch.vtensor<[?,?],f32>, !torch.int -> !torch.vtensor<[?,?],f32>
return %0 : !torch.vtensor<[?,?],f32>
}
// -----
[MHLO] support non-constant torch scalar in BasicOps (#1134) 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
2022-08-03 08:16:31 +08:00
// CHECK-LABEL: func.func @torch.aten.subtensor$alpha(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?],f32>, %[[ARG1:.*]]: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?],f32> -> tensor<?x?xf32>
// CHECK: %[[T1:.*]] = torch_c.to_builtin_tensor %[[ARG1]] : !torch.vtensor<[?,?],f32> -> tensor<?x?xf32>
// CHECK: %[[INT2:.*]] = torch.constant.int 2
// CHECK: %[[T2:.*]] = torch_c.to_i64 %[[INT2]]
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]] : tensor<1xi64>
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: %[[T3:.*]] = stablehlo.convert %[[FROM_ELEMENTS]] : (tensor<1xi64>) -> tensor<1xf32>
// CHECK: %[[T4:.*]] = stablehlo.reshape %[[T3]] : (tensor<1xf32>) -> tensor<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
// CHECK: %[[T5:.*]] = chlo.broadcast_multiply %[[T1]], %[[T4]] : (tensor<?x?xf32>, tensor<f32>) -> tensor<?x?xf32>
// CHECK: %[[T6:.*]] = chlo.broadcast_subtract %[[T0]], %[[T5]] : (tensor<?x?xf32>, tensor<?x?xf32>) -> tensor<?x?xf32>
// CHECK: %[[T7:.*]] = torch_c.from_builtin_tensor %[[T6]] : tensor<?x?xf32> -> !torch.vtensor<[?,?],f32>
// CHECK: return %[[T7]] : !torch.vtensor<[?,?],f32>
[MHLO] add aten::gelu op pattern (#1127) add aten::gelu op pattern, and moved some unit tests from basic.mlir to elementwise.mlir
2022-08-02 15:01:30 +08:00
func.func @torch.aten.subtensor$alpha(%arg0: !torch.vtensor<[?,?],f32>, %arg1: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
%int2 = torch.constant.int 2
%0 = torch.aten.sub.Tensor %arg0, %arg1, %int2 : !torch.vtensor<[?,?],f32>, !torch.vtensor<[?,?],f32>, !torch.int -> !torch.vtensor<[?,?],f32>
return %0 : !torch.vtensor<[?,?],f32>
}
// -----
[MHLO] support non-constant torch scalar in BasicOps (#1134) 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
2022-08-03 08:16:31 +08:00
// CHECK-LABEL: func.func @torch.aten.subtensor$promote(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?],si32>, %[[ARG1:.*]]: !torch.vtensor<[?,?],si64>) -> !torch.vtensor<[?,?],si64> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?],si32> -> tensor<?x?xi32>
// CHECK: %[[T1:.*]] = torch_c.to_builtin_tensor %[[ARG1]] : !torch.vtensor<[?,?],si64> -> tensor<?x?xi64>
// CHECK: %[[INT1:.*]] = torch.constant.int 1
mhlo: migrate conversion to stablehlo (#1840) This patch replaces all MHLO operations with their StableHLO counterparts and adds a validation pass to ensure that no MHLO operations remain before translating all Stablehlo operations to the MHLO dialect for further lowering to the Linalg dialect. This patch also updates all lit tests so that they refer to the `convert-torch-to-stablehlo` pass and so that they check for StableHLO operations.
2023-02-02 21:29:47 +08:00
// CHECK: %[[T2:.*]] = stablehlo.convert %[[T0]] : (tensor<?x?xi32>) -> tensor<?x?xi64>
[MHLO] support non-constant torch scalar in BasicOps (#1134) See RFC https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-08-03 08:16:31 +08:00
// CHECK: %[[T3:.*]] = chlo.broadcast_subtract %[[T2]], %[[T1]] : (tensor<?x?xi64>, tensor<?x?xi64>) -> tensor<?x?xi64>
// CHECK: %[[T4:.*]] = torch_c.from_builtin_tensor %[[T3]] : tensor<?x?xi64> -> !torch.vtensor<[?,?],si64>
// CHECK: return %[[T4]] : !torch.vtensor<[?,?],si64>
[MHLO] add aten::gelu op pattern (#1127) add aten::gelu op pattern, and moved some unit tests from basic.mlir to elementwise.mlir
2022-08-02 15:01:30 +08:00
func.func @torch.aten.subtensor$promote(%arg0: !torch.vtensor<[?,?],si32>, %arg1: !torch.vtensor<[?,?],si64>) -> !torch.vtensor<[?,?],si64> {
%int1 = torch.constant.int 1
%0 = torch.aten.sub.Tensor %arg0, %arg1, %int1 : !torch.vtensor<[?,?],si32>, !torch.vtensor<[?,?],si64>, !torch.int -> !torch.vtensor<[?,?],si64>
return %0 : !torch.vtensor<[?,?],si64>
}
// -----
[MHLO] support non-constant torch scalar in BasicOps (#1134) 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
2022-08-03 08:16:31 +08:00
// CHECK-LABEL: func.func @torch.aten.mulscalar$basic(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?],f32> -> tensor<?x?xf32>
// CHECK: %[[INT9:.*]] = torch.constant.int 9
// CHECK: %[[T1:.*]] = torch_c.to_i64 %[[INT9]]
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]] : tensor<1xi64>
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.convert %[[FROM_ELEMENTS]] : (tensor<1xi64>) -> tensor<1xf32>
// CHECK: %[[T3:.*]] = stablehlo.reshape %[[T2]] : (tensor<1xf32>) -> tensor<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
// CHECK: %[[T4:.*]] = chlo.broadcast_multiply %[[T0]], %[[T3]] : (tensor<?x?xf32>, tensor<f32>) -> tensor<?x?xf32>
// CHECK: %[[T5:.*]] = torch_c.from_builtin_tensor %[[T4]] : tensor<?x?xf32> -> !torch.vtensor<[?,?],f32>
// CHECK: return %[[T5]] : !torch.vtensor<[?,?],f32>
[MHLO] add aten::gelu op pattern (#1127) add aten::gelu op pattern, and moved some unit tests from basic.mlir to elementwise.mlir
2022-08-02 15:01:30 +08:00
func.func @torch.aten.mulscalar$basic(%arg0: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
%int9 = torch.constant.int 9
%0 = torch.aten.mul.Scalar %arg0, %int9 : !torch.vtensor<[?,?],f32>, !torch.int -> !torch.vtensor<[?,?],f32>
return %0 : !torch.vtensor<[?,?],f32>
}
// -----
[MHLO] support non-constant torch scalar in BasicOps (#1134) 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
2022-08-03 08:16:31 +08:00
// CHECK-LABEL: func.func @torch.aten.multensor$basic(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?],f32>, %[[ARG1:.*]]: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?],f32> -> tensor<?x?xf32>
// CHECK: %[[T1:.*]] = torch_c.to_builtin_tensor %[[ARG1]] : !torch.vtensor<[?,?],f32> -> tensor<?x?xf32>
// CHECK: %[[T2:.*]] = chlo.broadcast_multiply %[[T0]], %[[T1]] : (tensor<?x?xf32>, tensor<?x?xf32>) -> tensor<?x?xf32>
// CHECK: %[[T3:.*]] = torch_c.from_builtin_tensor %[[T2]] : tensor<?x?xf32> -> !torch.vtensor<[?,?],f32>
// CHECK: return %[[T3]] : !torch.vtensor<[?,?],f32>
[MHLO] add aten::gelu op pattern (#1127) add aten::gelu op pattern, and moved some unit tests from basic.mlir to elementwise.mlir
2022-08-02 15:01:30 +08:00
func.func @torch.aten.multensor$basic(%arg0: !torch.vtensor<[?,?],f32>, %arg1: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
%0 = torch.aten.mul.Tensor %arg0, %arg1 : !torch.vtensor<[?,?],f32>, !torch.vtensor<[?,?],f32> -> !torch.vtensor<[?,?],f32>
return %0 : !torch.vtensor<[?,?],f32>
}
// -----
[MHLO] support non-constant torch scalar in BasicOps (#1134) 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
2022-08-03 08:16:31 +08:00
// CHECK-LABEL: func.func @torch.aten.divscalar$basic(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?],f32> -> tensor<?x?xf32>
// CHECK: %[[INT9:.*]] = torch.constant.int 9
// CHECK: %[[T1:.*]] = torch_c.to_i64 %[[INT9]]
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]] : tensor<1xi64>
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.convert %[[FROM_ELEMENTS]] : (tensor<1xi64>) -> tensor<1xf32>
// CHECK: %[[T3:.*]] = stablehlo.reshape %[[T2]] : (tensor<1xf32>) -> tensor<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
// CHECK: %[[T4:.*]] = chlo.broadcast_divide %[[T0]], %[[T3]] : (tensor<?x?xf32>, tensor<f32>) -> tensor<?x?xf32>
// CHECK: %[[T5:.*]] = torch_c.from_builtin_tensor %[[T4]] : tensor<?x?xf32> -> !torch.vtensor<[?,?],f32>
// CHECK: return %[[T5]] : !torch.vtensor<[?,?],f32>
[MHLO] add aten::gelu op pattern (#1127) add aten::gelu op pattern, and moved some unit tests from basic.mlir to elementwise.mlir
2022-08-02 15:01:30 +08:00
func.func @torch.aten.divscalar$basic(%arg0: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
%int9 = torch.constant.int 9
%0 = torch.aten.div.Scalar %arg0, %int9 : !torch.vtensor<[?,?],f32>, !torch.int -> !torch.vtensor<[?,?],f32>
return %0 : !torch.vtensor<[?,?],f32>
}
// -----
[MHLO] support non-constant torch scalar in BasicOps (#1134) 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
2022-08-03 08:16:31 +08:00
// CHECK-LABEL: func.func @torch.aten.divtensor$basic(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?],f32>, %[[ARG1:.*]]: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?],f32> -> tensor<?x?xf32>
// CHECK: %[[T1:.*]] = torch_c.to_builtin_tensor %[[ARG1]] : !torch.vtensor<[?,?],f32> -> tensor<?x?xf32>
// CHECK: %[[T2:.*]] = chlo.broadcast_divide %[[T0]], %[[T1]] : (tensor<?x?xf32>, tensor<?x?xf32>) -> tensor<?x?xf32>
// CHECK: %[[T3:.*]] = torch_c.from_builtin_tensor %[[T2]] : tensor<?x?xf32> -> !torch.vtensor<[?,?],f32>
// CHECK: return %[[T3]] : !torch.vtensor<[?,?],f32>
[MHLO] add aten::gelu op pattern (#1127) add aten::gelu op pattern, and moved some unit tests from basic.mlir to elementwise.mlir
2022-08-02 15:01:30 +08:00
func.func @torch.aten.divtensor$basic(%arg0: !torch.vtensor<[?,?],f32>, %arg1: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
%0 = torch.aten.div.Tensor %arg0, %arg1 : !torch.vtensor<[?,?],f32>, !torch.vtensor<[?,?],f32> -> !torch.vtensor<[?,?],f32>
return %0 : !torch.vtensor<[?,?],f32>
}
// -----
[MHLO] support non-constant torch scalar in BasicOps (#1134) 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
2022-08-03 08:16:31 +08:00
// CHECK-LABEL: func.func @torch.aten.gt.scalar(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],i1> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?],f32> -> tensor<?x?xf32>
// CHECK: %[[INT3:.*]] = torch.constant.int 3
// CHECK: %[[T1:.*]] = torch_c.to_i64 %[[INT3]]
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]] : tensor<1xi64>
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.convert %[[FROM_ELEMENTS]] : (tensor<1xi64>) -> tensor<1xf32>
// CHECK: %[[T3:.*]] = stablehlo.reshape %[[T2]] : (tensor<1xf32>) -> tensor<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
// CHECK: %[[T4:.*]] = chlo.broadcast_compare %[[T0]], %[[T3]] {compare_type = #chlo<comparison_type FLOAT>, comparison_direction = #chlo<comparison_direction GT>} : (tensor<?x?xf32>, tensor<f32>) -> tensor<?x?xi1>
// CHECK: %[[T5:.*]] = torch_c.from_builtin_tensor %[[T4]] : tensor<?x?xi1> -> !torch.vtensor<[?,?],i1>
// CHECK: return %[[T5]] : !torch.vtensor<[?,?],i1>
[MHLO] add aten::gelu op pattern (#1127) add aten::gelu op pattern, and moved some unit tests from basic.mlir to elementwise.mlir
2022-08-02 15:01:30 +08:00
func.func @torch.aten.gt.scalar(%arg0: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],i1> {
%int3 = torch.constant.int 3
%0 = torch.aten.gt.Scalar %arg0, %int3 : !torch.vtensor<[?,?],f32>, !torch.int -> !torch.vtensor<[?,?],i1>
return %0 : !torch.vtensor<[?,?],i1>
}
// -----
[MHLO] support non-constant torch scalar in BasicOps (#1134) 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
2022-08-03 08:16:31 +08:00
// CHECK-LABEL: func.func @torch.aten.gt.tensor(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?],f32>, %[[ARG1:.*]]: !torch.vtensor<[64],f32>) -> !torch.vtensor<[?,?],i1> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?],f32> -> tensor<?x?xf32>
// CHECK: %[[T1:.*]] = torch_c.to_builtin_tensor %[[ARG1]] : !torch.vtensor<[64],f32> -> tensor<64xf32>
build: update llvm tag to 00d648bd (#1307) - Update MHLO commit to build with LLVM commit hash 00d648bd - Update TorchToMhlo code to work with Stablehlo - Re-enabled two failing TOSA tests, thus resolving Github Issue #1231
2022-08-31 03:44:00 +08:00
// CHECK: %[[T2:.*]] = chlo.broadcast_compare %[[T0]], %[[T1]] {compare_type = #chlo<comparison_type FLOAT>, comparison_direction = #chlo<comparison_direction GT>} : (tensor<?x?xf32>, tensor<64xf32>) -> tensor<?x?xi1>
[MHLO] support non-constant torch scalar in BasicOps (#1134) 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
2022-08-03 08:16:31 +08:00
// CHECK: %[[T3:.*]] = torch_c.from_builtin_tensor %[[T2]] : tensor<?x?xi1> -> !torch.vtensor<[?,?],i1>
// CHECK: return %[[T3]] : !torch.vtensor<[?,?],i1>
[MHLO] add aten::gelu op pattern (#1127) add aten::gelu op pattern, and moved some unit tests from basic.mlir to elementwise.mlir
2022-08-02 15:01:30 +08:00
func.func @torch.aten.gt.tensor(%arg0: !torch.vtensor<[?,?],f32>, %arg1: !torch.vtensor<[64],f32>) -> !torch.vtensor<[?,?],i1> {
%0 = torch.aten.gt.Tensor %arg0, %arg1 : !torch.vtensor<[?,?],f32>, !torch.vtensor<[64],f32> -> !torch.vtensor<[?,?],i1>
return %0 : !torch.vtensor<[?,?],i1>
}
// -----
[MHLO] support non-constant torch scalar in BasicOps (#1134) 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
2022-08-03 08:16:31 +08:00
// CHECK-LABEL: func.func @torch.aten.lt.tensor(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?],f32>, %[[ARG1:.*]]: !torch.vtensor<[64],f32>) -> !torch.vtensor<[?,?],i1> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?],f32> -> tensor<?x?xf32>
// CHECK: %[[T1:.*]] = torch_c.to_builtin_tensor %[[ARG1]] : !torch.vtensor<[64],f32> -> tensor<64xf32>
build: update llvm tag to 00d648bd (#1307) - Update MHLO commit to build with LLVM commit hash 00d648bd - Update TorchToMhlo code to work with Stablehlo - Re-enabled two failing TOSA tests, thus resolving Github Issue #1231
2022-08-31 03:44:00 +08:00
// CHECK: %[[T2:.*]] = chlo.broadcast_compare %[[T0]], %[[T1]] {compare_type = #chlo<comparison_type FLOAT>, comparison_direction = #chlo<comparison_direction LT>} : (tensor<?x?xf32>, tensor<64xf32>) -> tensor<?x?xi1>
[MHLO] support non-constant torch scalar in BasicOps (#1134) 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
2022-08-03 08:16:31 +08:00
// CHECK: %[[T3:.*]] = torch_c.from_builtin_tensor %[[T2]] : tensor<?x?xi1> -> !torch.vtensor<[?,?],i1>
// CHECK: return %[[T3]] : !torch.vtensor<[?,?],i1>
[MHLO] add aten::gelu op pattern (#1127) add aten::gelu op pattern, and moved some unit tests from basic.mlir to elementwise.mlir
2022-08-02 15:01:30 +08:00
func.func @torch.aten.lt.tensor(%arg0: !torch.vtensor<[?,?],f32>, %arg1: !torch.vtensor<[64],f32>) -> !torch.vtensor<[?,?],i1> {
%0 = torch.aten.lt.Tensor %arg0, %arg1 : !torch.vtensor<[?,?],f32>, !torch.vtensor<[64],f32> -> !torch.vtensor<[?,?],i1>
return %0 : !torch.vtensor<[?,?],i1>
}
// -----
[MHLO] support non-constant torch scalar in BasicOps (#1134) 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
2022-08-03 08:16:31 +08:00
// CHECK-LABEL: func.func @torch.aten.eq.tensor(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?],f32>, %[[ARG1:.*]]: !torch.vtensor<[64],f32>) -> !torch.vtensor<[?,?],i1> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?],f32> -> tensor<?x?xf32>
// CHECK: %[[T1:.*]] = torch_c.to_builtin_tensor %[[ARG1]] : !torch.vtensor<[64],f32> -> tensor<64xf32>
build: update llvm tag to 00d648bd (#1307) - Update MHLO commit to build with LLVM commit hash 00d648bd - Update TorchToMhlo code to work with Stablehlo - Re-enabled two failing TOSA tests, thus resolving Github Issue #1231
2022-08-31 03:44:00 +08:00
// CHECK: %[[T2:.*]] = chlo.broadcast_compare %[[T0]], %[[T1]] {compare_type = #chlo<comparison_type FLOAT>, comparison_direction = #chlo<comparison_direction EQ>} : (tensor<?x?xf32>, tensor<64xf32>) -> tensor<?x?xi1>
[MHLO] support non-constant torch scalar in BasicOps (#1134) 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
2022-08-03 08:16:31 +08:00
// CHECK: %[[T3:.*]] = torch_c.from_builtin_tensor %[[T2]] : tensor<?x?xi1> -> !torch.vtensor<[?,?],i1>
// CHECK: return %[[T3]] : !torch.vtensor<[?,?],i1>
[MHLO] add aten::gelu op pattern (#1127) add aten::gelu op pattern, and moved some unit tests from basic.mlir to elementwise.mlir
2022-08-02 15:01:30 +08:00
func.func @torch.aten.eq.tensor(%arg0: !torch.vtensor<[?,?],f32>, %arg1: !torch.vtensor<[64],f32>) -> !torch.vtensor<[?,?],i1> {
%0 = torch.aten.eq.Tensor %arg0, %arg1 : !torch.vtensor<[?,?],f32>, !torch.vtensor<[64],f32> -> !torch.vtensor<[?,?],i1>
return %0 : !torch.vtensor<[?,?],i1>
}
// -----
[MHLO] support non-constant torch scalar in BasicOps (#1134) 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
2022-08-03 08:16:31 +08:00
// CHECK-LABEL: func.func @torch.aten.ne.tensor(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?],f32>, %[[ARG1:.*]]: !torch.vtensor<[64],f32>) -> !torch.vtensor<[?,?],i1> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?],f32> -> tensor<?x?xf32>
// CHECK: %[[T1:.*]] = torch_c.to_builtin_tensor %[[ARG1]] : !torch.vtensor<[64],f32> -> tensor<64xf32>
build: update llvm tag to 00d648bd (#1307) - Update MHLO commit to build with LLVM commit hash 00d648bd - Update TorchToMhlo code to work with Stablehlo - Re-enabled two failing TOSA tests, thus resolving Github Issue #1231
2022-08-31 03:44:00 +08:00
// CHECK: %[[T2:.*]] = chlo.broadcast_compare %[[T0]], %[[T1]] {compare_type = #chlo<comparison_type FLOAT>, comparison_direction = #chlo<comparison_direction NE>} : (tensor<?x?xf32>, tensor<64xf32>) -> tensor<?x?xi1>
[MHLO] support non-constant torch scalar in BasicOps (#1134) 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
2022-08-03 08:16:31 +08:00
// CHECK: %[[T3:.*]] = torch_c.from_builtin_tensor %[[T2]] : tensor<?x?xi1> -> !torch.vtensor<[?,?],i1>
// CHECK: return %[[T3]] : !torch.vtensor<[?,?],i1>
[MHLO] add aten::gelu op pattern (#1127) add aten::gelu op pattern, and moved some unit tests from basic.mlir to elementwise.mlir
2022-08-02 15:01:30 +08:00
func.func @torch.aten.ne.tensor(%arg0: !torch.vtensor<[?,?],f32>, %arg1: !torch.vtensor<[64],f32>) -> !torch.vtensor<[?,?],i1> {
%0 = torch.aten.ne.Tensor %arg0, %arg1 : !torch.vtensor<[?,?],f32>, !torch.vtensor<[64],f32> -> !torch.vtensor<[?,?],i1>
return %0 : !torch.vtensor<[?,?],i1>
}
// -----
[MHLO] support non-constant torch scalar in BasicOps (#1134) 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
2022-08-03 08:16:31 +08:00
// CHECK-LABEL: func.func @torch.aten.permute$basic(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[4,64],f32>) -> !torch.vtensor<[64,4],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[4,64],f32> -> tensor<4x64xf32>
// CHECK: %[[INT0:.*]] = torch.constant.int 0
// CHECK: %[[INT1:.*]] = torch.constant.int 1
// CHECK: %[[T1:.*]] = torch.prim.ListConstruct %[[INT1]], %[[INT0]] : (!torch.int, !torch.int) -> !torch.list<int>
mhlo: migrate conversion to stablehlo (#1840) This patch replaces all MHLO operations with their StableHLO counterparts and adds a validation pass to ensure that no MHLO operations remain before translating all Stablehlo operations to the MHLO dialect for further lowering to the Linalg dialect. This patch also updates all lit tests so that they refer to the `convert-torch-to-stablehlo` pass and so that they check for StableHLO operations.
2023-02-02 21:29:47 +08:00
// CHECK: %[[T2:.*]] = stablehlo.transpose %[[T0]], dims = [1, 0] : (tensor<4x64xf32>) -> tensor<64x4xf32>
[MHLO] support non-constant torch scalar in BasicOps (#1134) See RFC https://github.com/llvm/torch-mlir/issues/999 Co-authored-by: Bairen Yi yibairen.byron@bytedance.com Co-authored-by: Jiawei Wu xremold@gmail.com Co-authored-by: Tianyou Guo tianyou.gty@alibaba-inc.com Co-authored-by: Xu Yan yancey.yx@alibaba-inc.com Co-authored-by: Ziheng Jiang ziheng.jiang@bytedance.com
2022-08-03 08:16:31 +08:00
// CHECK: %[[T3:.*]] = torch_c.from_builtin_tensor %[[T2]] : tensor<64x4xf32> -> !torch.vtensor<[64,4],f32>
// CHECK: return %[[T3]] : !torch.vtensor<[64,4],f32>
[MHLO] add aten::gelu op pattern (#1127) add aten::gelu op pattern, and moved some unit tests from basic.mlir to elementwise.mlir
2022-08-02 15:01:30 +08:00
func.func @torch.aten.permute$basic(%arg0: !torch.vtensor<[4,64],f32>) -> !torch.vtensor<[64,4],f32> {
%int0 = torch.constant.int 0
%int1 = torch.constant.int 1
%0 = torch.prim.ListConstruct %int1, %int0 : (!torch.int, !torch.int) -> !torch.list<int>
%1 = torch.aten.permute %arg0, %0 : !torch.vtensor<[4,64],f32>, !torch.list<int> -> !torch.vtensor<[64,4],f32>
return %1 : !torch.vtensor<[64,4],f32>
}
// -----
[MHLO] support non-constant torch scalar in BasicOps (#1134) 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
2022-08-03 08:16:31 +08:00
// CHECK-LABEL: func.func @torch.aten.relu(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?],f32> -> tensor<?x?xf32>
// CHECK: %[[T1:.*]] = "chlo.constant_like"(%[[T0]]) {value = 0.000000e+00 : f32} : (tensor<?x?xf32>) -> tensor<?x?xf32>
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.maximum %[[T0]], %[[T1]] : tensor<?x?xf32>
[MHLO] support non-constant torch scalar in BasicOps (#1134) 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
2022-08-03 08:16:31 +08:00
// CHECK: %[[T3:.*]] = torch_c.from_builtin_tensor %[[T2]] : tensor<?x?xf32> -> !torch.vtensor<[?,?],f32>
// CHECK: return %[[T3]] : !torch.vtensor<[?,?],f32>
[MHLO] add aten::gelu op pattern (#1127) add aten::gelu op pattern, and moved some unit tests from basic.mlir to elementwise.mlir
2022-08-02 15:01:30 +08:00
func.func @torch.aten.relu(%arg0: !torch.vtensor<[?,?],f32>) -> !torch.vtensor<[?,?],f32> {
%0 = torch.aten.relu %arg0 : !torch.vtensor<[?,?],f32> -> !torch.vtensor<[?,?],f32>
return %0 : !torch.vtensor<[?,?],f32>
}
[MHLO] support non-constant torch scalar in BasicOps (#1134) 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
2022-08-03 08:16:31 +08:00
// -----
// CHECK-LABEL: func.func @torch.aten.addscalar$variable(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?],f32>, %[[ARG1:.*]]: !torch.float) -> !torch.vtensor<[?,?],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?],f32> -> tensor<?x?xf32>
// CHECK: %[[T1:.*]] = torch_c.to_f64 %[[ARG1]]
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]] : tensor<1xf64>
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.convert %[[FROM_ELEMENTS]] : (tensor<1xf64>) -> tensor<1xf32>
// CHECK: %[[T3:.*]] = stablehlo.reshape %[[T2]] : (tensor<1xf32>) -> tensor<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
// CHECK: %[[FROM_ELEMENTS_0:.*]] = tensor.from_elements %[[T1]] : tensor<1xf64>
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.convert %[[FROM_ELEMENTS_0]] : (tensor<1xf64>) -> tensor<1xf32>
// CHECK: %[[T5:.*]] = stablehlo.reshape %[[T4]] : (tensor<1xf32>) -> tensor<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
// CHECK: %[[T6:.*]] = chlo.broadcast_multiply %[[T3]], %[[T5]] : (tensor<f32>, tensor<f32>) -> tensor<f32>
// CHECK: %[[T7:.*]] = chlo.broadcast_add %[[T0]], %[[T6]] : (tensor<?x?xf32>, tensor<f32>) -> tensor<?x?xf32>
// CHECK: %[[T8:.*]] = torch_c.from_builtin_tensor %[[T7]] : tensor<?x?xf32> -> !torch.vtensor<[?,?],f32>
// CHECK: return %[[T8]] : !torch.vtensor<[?,?],f32>
[MHLO] support non-constant torch scalar in BasicOps (#1134) 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
2022-08-03 08:16:31 +08:00
func.func @torch.aten.addscalar$variable(%arg0: !torch.vtensor<[?,?],f32>, %arg1: !torch.float) -> !torch.vtensor<[?,?],f32> {
%0 = torch.aten.add.Scalar %arg0, %arg1, %arg1: !torch.vtensor<[?,?],f32>, !torch.float, !torch.float -> !torch.vtensor<[?,?],f32>
return %0 : !torch.vtensor<[?,?],f32>
}
// -----
// CHECK-LABEL: func.func @torch.aten.addtensor$variable(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?],f32>, %[[ARG1:.*]]: !torch.vtensor<[?,?],f32>, %[[ARG2:.*]]: !torch.float) -> !torch.vtensor<[?,?],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?],f32> -> tensor<?x?xf32>
// CHECK: %[[T1:.*]] = torch_c.to_builtin_tensor %[[ARG1]] : !torch.vtensor<[?,?],f32> -> tensor<?x?xf32>
// CHECK: %[[T2:.*]] = torch_c.to_f64 %[[ARG2]]
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]] : tensor<1xf64>
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: %[[T3:.*]] = stablehlo.convert %[[FROM_ELEMENTS]] : (tensor<1xf64>) -> tensor<1xf32>
// CHECK: %[[T4:.*]] = stablehlo.reshape %[[T3]] : (tensor<1xf32>) -> tensor<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
// CHECK: %[[T5:.*]] = chlo.broadcast_multiply %[[T1]], %[[T4]] : (tensor<?x?xf32>, tensor<f32>) -> tensor<?x?xf32>
// CHECK: %[[T6:.*]] = chlo.broadcast_add %[[T0]], %[[T5]] : (tensor<?x?xf32>, tensor<?x?xf32>) -> tensor<?x?xf32>
// CHECK: %[[T7:.*]] = torch_c.from_builtin_tensor %[[T6]] : tensor<?x?xf32> -> !torch.vtensor<[?,?],f32>
// CHECK: return %[[T7]] : !torch.vtensor<[?,?],f32>
[MHLO] support non-constant torch scalar in BasicOps (#1134) 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
2022-08-03 08:16:31 +08:00
func.func @torch.aten.addtensor$variable(%arg0: !torch.vtensor<[?,?],f32>, %arg1: !torch.vtensor<[?,?],f32>, %arg2: !torch.float) -> !torch.vtensor<[?,?],f32> {
%0 = torch.aten.add.Tensor %arg0, %arg1, %arg2: !torch.vtensor<[?,?],f32>, !torch.vtensor<[?,?],f32>, !torch.float -> !torch.vtensor<[?,?],f32>
return %0 : !torch.vtensor<[?,?],f32>
}
// -----
// CHECK-LABEL: func.func @torch.aten.mulscalar$variable(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?],f32>, %[[ARG1:.*]]: !torch.int) -> !torch.vtensor<[?,?],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?],f32> -> tensor<?x?xf32>
// CHECK: %[[T1:.*]] = torch_c.to_i64 %[[ARG1]]
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]] : tensor<1xi64>
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.convert %[[FROM_ELEMENTS]] : (tensor<1xi64>) -> tensor<1xf32>
// CHECK: %[[T3:.*]] = stablehlo.reshape %[[T2]] : (tensor<1xf32>) -> tensor<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
// CHECK: %[[T4:.*]] = chlo.broadcast_multiply %[[T0]], %[[T3]] : (tensor<?x?xf32>, tensor<f32>) -> tensor<?x?xf32>
// CHECK: %[[T5:.*]] = torch_c.from_builtin_tensor %[[T4]] : tensor<?x?xf32> -> !torch.vtensor<[?,?],f32>
// CHECK: return %[[T5]] : !torch.vtensor<[?,?],f32>
[MHLO] support non-constant torch scalar in BasicOps (#1134) 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
2022-08-03 08:16:31 +08:00
func.func @torch.aten.mulscalar$variable(%arg0: !torch.vtensor<[?,?],f32>, %arg1: !torch.int) -> !torch.vtensor<[?,?],f32> {
%0 = torch.aten.mul.Scalar %arg0, %arg1: !torch.vtensor<[?,?],f32>, !torch.int -> !torch.vtensor<[?,?],f32>
return %0 : !torch.vtensor<[?,?],f32>
}
// -----
// CHECK-LABEL: func.func @torch.aten.divscalar$variable(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?],f32>, %[[ARG1:.*]]: !torch.int) -> !torch.vtensor<[?,?],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?],f32> -> tensor<?x?xf32>
// CHECK: %[[T1:.*]] = torch_c.to_i64 %[[ARG1]]
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]] : tensor<1xi64>
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.convert %[[FROM_ELEMENTS]] : (tensor<1xi64>) -> tensor<1xf32>
// CHECK: %[[T3:.*]] = stablehlo.reshape %[[T2]] : (tensor<1xf32>) -> tensor<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
// CHECK: %[[T4:.*]] = chlo.broadcast_divide %[[T0]], %[[T3]] : (tensor<?x?xf32>, tensor<f32>) -> tensor<?x?xf32>
// CHECK: %[[T5:.*]] = torch_c.from_builtin_tensor %[[T4]] : tensor<?x?xf32> -> !torch.vtensor<[?,?],f32>
// CHECK: return %[[T5]] : !torch.vtensor<[?,?],f32>
[MHLO] support non-constant torch scalar in BasicOps (#1134) 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
2022-08-03 08:16:31 +08:00
func.func @torch.aten.divscalar$variable(%arg0: !torch.vtensor<[?,?],f32>, %arg1: !torch.int) -> !torch.vtensor<[?,?],f32> {
%0 = torch.aten.div.Scalar %arg0, %arg1: !torch.vtensor<[?,?],f32>, !torch.int -> !torch.vtensor<[?,?],f32>
return %0 : !torch.vtensor<[?,?],f32>
}
// -----
// CHECK-LABEL: func.func @torch.aten.gt.scalar$variable(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?],f32>, %[[ARG1:.*]]: !torch.int) -> !torch.vtensor<[?,?],i1> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?],f32> -> tensor<?x?xf32>
// CHECK: %[[T1:.*]] = torch_c.to_i64 %[[ARG1]]
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]] : tensor<1xi64>
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.convert %[[FROM_ELEMENTS]] : (tensor<1xi64>) -> tensor<1xf32>
// CHECK: %[[T3:.*]] = stablehlo.reshape %[[T2]] : (tensor<1xf32>) -> tensor<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
// CHECK: %[[T4:.*]] = chlo.broadcast_compare %[[T0]], %[[T3]] {compare_type = #chlo<comparison_type FLOAT>, comparison_direction = #chlo<comparison_direction GT>} : (tensor<?x?xf32>, tensor<f32>) -> tensor<?x?xi1>
// CHECK: %[[T5:.*]] = torch_c.from_builtin_tensor %[[T4]] : tensor<?x?xi1> -> !torch.vtensor<[?,?],i1>
// CHECK: return %[[T5]] : !torch.vtensor<[?,?],i1>
[MHLO] support non-constant torch scalar in BasicOps (#1134) 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
2022-08-03 08:16:31 +08:00
func.func @torch.aten.gt.scalar$variable(%arg0: !torch.vtensor<[?,?],f32>, %arg1: !torch.int) -> !torch.vtensor<[?,?],i1> {
%0 = torch.aten.gt.Scalar %arg0, %arg1: !torch.vtensor<[?,?],f32>, !torch.int -> !torch.vtensor<[?,?],i1>
return %0 : !torch.vtensor<[?,?],i1>
}
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] fix tensor mode aten.div op pattern (#1160) * [MHLO] fix tensor mode aten.div op pattern See RFC #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>
2022-08-06 23:38:06 +08:00
// CHECK-LABEL: func.func @torch.aten.div.Tensor_mode$trunc(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?,?,?],f32>, %[[ARG1:.*]]: !torch.vtensor<[?,?,?,?],f32>) -> !torch.vtensor<[?,?,?,?],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?,?,?],f32> -> tensor<?x?x?x?xf32>
// CHECK: %[[T1:.*]] = torch_c.to_builtin_tensor %[[ARG1]] : !torch.vtensor<[?,?,?,?],f32> -> tensor<?x?x?x?xf32>
// CHECK: %[[STR:.*]] = torch.constant.str "trunc"
// CHECK: %[[T2:.*]] = chlo.broadcast_divide %[[T0]], %[[T1]] : (tensor<?x?x?x?xf32>, tensor<?x?x?x?xf32>) -> tensor<?x?x?x?xf32>
mhlo: migrate conversion to stablehlo (#1840) This patch replaces all MHLO operations with their StableHLO counterparts and adds a validation pass to ensure that no MHLO operations remain before translating all Stablehlo operations to the MHLO dialect for further lowering to the Linalg dialect. This patch also updates all lit tests so that they refer to the `convert-torch-to-stablehlo` pass and so that they check for StableHLO operations.
2023-02-02 21:29:47 +08:00
// CHECK: %[[T3:.*]] = stablehlo.sign %[[T2]] : tensor<?x?x?x?xf32>
// CHECK: %[[T4:.*]] = stablehlo.abs %[[T2]] : tensor<?x?x?x?xf32>
// CHECK: %[[T5:.*]] = stablehlo.floor %[[T4]] : tensor<?x?x?x?xf32>
// CHECK: %[[T6:.*]] = stablehlo.multiply %[[T3]], %[[T5]] : tensor<?x?x?x?xf32>
[MHLO] fix tensor mode aten.div op pattern (#1160) * [MHLO] fix tensor mode aten.div op pattern See RFC #999 Co-authored-by: Bairen Yi <yibairen.byron@bytedance.com> Co-authored-by: Jiawei Wu <xremold@gmail.com> Co-authored-by: Tianyou Guo <tianyou.gty@alibaba-inc.com> Co-authored-by: Xu Yan <yancey.yx@alibaba-inc.com> Co-authored-by: Ziheng Jiang <ziheng.jiang@bytedance.com>
2022-08-06 23:38:06 +08:00
// CHECK: %[[T7:.*]] = torch_c.from_builtin_tensor %[[T6]] : tensor<?x?x?x?xf32> -> !torch.vtensor<[?,?,?,?],f32>
// CHECK: return %[[T7]] : !torch.vtensor<[?,?,?,?],f32>
func.func @torch.aten.div.Tensor_mode$trunc(%arg0: !torch.vtensor<[?,?,?,?],f32>, %arg1: !torch.vtensor<[?,?,?,?],f32>) -> !torch.vtensor<[?,?,?,?],f32> {
%str = torch.constant.str "trunc"
%0 = torch.aten.div.Tensor_mode %arg0, %arg1, %str : !torch.vtensor<[?,?,?,?],f32>, !torch.vtensor<[?,?,?,?],f32>, !torch.str -> !torch.vtensor<[?,?,?,?],f32>
return %0 : !torch.vtensor<[?,?,?,?],f32>
}
// -----
// CHECK-LABEL: func.func @torch.aten.div.Tensor_mode$floor(
// CHECK-SAME: %[[ARG0:.*]]: !torch.vtensor<[?,?,?,?],f32>, %[[ARG1:.*]]: !torch.vtensor<[?,?,?,?],f32>) -> !torch.vtensor<[?,?,?,?],f32> {
// CHECK: %[[T0:.*]] = torch_c.to_builtin_tensor %[[ARG0]] : !torch.vtensor<[?,?,?,?],f32> -> tensor<?x?x?x?xf32>
// CHECK: %[[T1:.*]] = torch_c.to_builtin_tensor %[[ARG1]] : !torch.vtensor<[?,?,?,?],f32> -> tensor<?x?x?x?xf32>
// CHECK: %[[STR:.*]] = torch.constant.str "floor"
// CHECK: %[[T2:.*]] = chlo.broadcast_divide %[[T0]], %[[T1]] : (tensor<?x?x?x?xf32>, tensor<?x?x?x?xf32>) -> tensor<?x?x?x?xf32>
mhlo: migrate conversion to stablehlo (#1840) This patch replaces all MHLO operations with their StableHLO counterparts and adds a validation pass to ensure that no MHLO operations remain before translating all Stablehlo operations to the MHLO dialect for further lowering to the Linalg dialect. This patch also updates all lit tests so that they refer to the `convert-torch-to-stablehlo` pass and so that they check for StableHLO operations.
2023-02-02 21:29:47 +08:00
// CHECK: %[[T3:.*]] = stablehlo.floor %[[T2]] : tensor<?x?x?x?xf32>
[MHLO] fix tensor mode aten.div op pattern (#1160) * [MHLO] fix tensor mode aten.div op pattern See RFC #999 Co-authored-by: Bairen Yi <yibairen.byron@bytedance.com> Co-authored-by: Jiawei Wu <xremold@gmail.com> Co-authored-by: Tianyou Guo <tianyou.gty@alibaba-inc.com> Co-authored-by: Xu Yan <yancey.yx@alibaba-inc.com> Co-authored-by: Ziheng Jiang <ziheng.jiang@bytedance.com>
2022-08-06 23:38:06 +08:00
// CHECK: %[[T4:.*]] = torch_c.from_builtin_tensor %[[T3]] : tensor<?x?x?x?xf32> -> !torch.vtensor<[?,?,?,?],f32>
// CHECK: return %[[T4]] : !torch.vtensor<[?,?,?,?],f32>
func.func @torch.aten.div.Tensor_mode$floor(%arg0: !torch.vtensor<[?,?,?,?],f32>, %arg1: !torch.vtensor<[?,?,?,?],f32>) -> !torch.vtensor<[?,?,?,?],f32> {
%str = torch.constant.str "floor"
%0 = torch.aten.div.Tensor_mode %arg0, %arg1, %str : !torch.vtensor<[?,?,?,?],f32>, !torch.vtensor<[?,?,?,?],f32>, !torch.str -> !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