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/TorchToLinalg/pooling.mlir

30 lines
2.0 KiB
MLIR
Raw Normal View History

Move TorchConversion dialect and TorchTo* into torch-mlir
2021-09-24 11:59:12 +08:00
// RUN: torch-mlir-opt <%s -convert-torch-to-linalg -split-input-file -verify-diagnostics | FileCheck %s
Add TorchToIREE and factor out TorchConversion dialect. This converts a basic list op (torch.prim.ListConstruct) to the IREE dialect. ``` def forward(self, x: float): return [x, x] ``` turns into: ``` builtin.func @forward(%arg0: !torch.float) -> !torch.list<!torch.float> { %0 = torch.prim.ListConstruct %arg0, %arg0 : (!torch.float, !torch.float) -> !torch.list<!torch.float> return %0 : !torch.list<!torch.float> } ``` which turns into: ``` builtin.func @forward(%arg0: f64) -> !iree.list<f64> { %c1 = constant 1 : index %c0 = constant 0 : index %c2 = constant 2 : index %0 = iree.list.create %c2 : !iree.list<f64> iree.list.set %0[%c0], %arg0 : !iree.list<f64>, f64 iree.list.set %0[%c1], %arg0 : !iree.list<f64>, f64 return %0 : !iree.list<f64> } ``` As part of doing this, I realized that it was time to formalize the IR form that we reach right before running TorchTo{Linalg,Std,...}. We now call it the "Torch backend contract". We then lower the "Torch backend contract" to the "npcomp backend contract", which involves the new TorchConversion (`torch_c`) dialect, which holds ops that need to operate on both the npcomp backend types (e.g. builtin tensors, i1, IREE list, etc.) and the `!torch` types. This made more sense, as I realized that if I didn't factor out `torch_c` then the Torch dialect would have a dependency on IREE dialect (we previously didn't notice this was an issue because we only depended on `builtin` types), which seemed wrong to me. Recommended review order: - TorchToIREE.cpp / `TorchToIREE/basic.mlir` - Look at the new structure of createTorchScriptToNpcompBackendPipeline. It now lives in TorchConversion/Transforms/Passes.cpp and cleanly calls into `Torch::createTorchScriptToTorchBackendPipeline` for the frontend lowering to the Torch backend contract. - Mechanical change extracting `torch_c.{to,from}_{i1,i64,f64,builtin_tensor,iree_list}` into a new TorchConversion dialect, and a few passes specific to the lowering from the Torch backend contract to the npcomp backend contract. - Minor fixes to TorchToLinalg.cpp to use unconverted operands (now that we convert lists as part of operand materialization, we need to use the original operands). Also added test for AtenMaxPool2dOp and fixed m_TorchConstantIntList. - TmpDeleteDeadIREELists pass. Temporary pass for deleting dead IREE lists that are created as part of operand materialization for conv/max pool/avg pool ops in TorchToLinalg.
2021-08-12 05:40:08 +08:00
Update llvm-project to 830c0b9023cd0cf91955900e0d96283e7a8c3711 - builder.getSymbolRefAttr is gone. - OpAsmOpInterface's getAsmResultNames method needs explicit override - a bunch of churn for builtin.func needing to be made explicit (and sometimes implicit?) - operation printers no longer need to print the operation name themselves. - snuck in beneficial trivial addition to TmpDeleteDeadIREEListsPass to test a particular upstream change e2e with my local patchset.
2021-09-04 02:38:00 +08:00
// CHECK-LABEL: func @forward
llvm: bump tag to e1318078 (#781) The updated LLVM code includes a patch to create bfloat16 array attributes, thus enabling a different patch to torch-mlir to flesh out support for the bfloat16 type.
2022-04-27 03:27:51 +08:00
func.func @forward(%arg0: !torch.vtensor<[?,?,?,?],f32>) -> !torch.vtensor<[?,?,?,?],f32> {
Add TorchToIREE and factor out TorchConversion dialect. This converts a basic list op (torch.prim.ListConstruct) to the IREE dialect. ``` def forward(self, x: float): return [x, x] ``` turns into: ``` builtin.func @forward(%arg0: !torch.float) -> !torch.list<!torch.float> { %0 = torch.prim.ListConstruct %arg0, %arg0 : (!torch.float, !torch.float) -> !torch.list<!torch.float> return %0 : !torch.list<!torch.float> } ``` which turns into: ``` builtin.func @forward(%arg0: f64) -> !iree.list<f64> { %c1 = constant 1 : index %c0 = constant 0 : index %c2 = constant 2 : index %0 = iree.list.create %c2 : !iree.list<f64> iree.list.set %0[%c0], %arg0 : !iree.list<f64>, f64 iree.list.set %0[%c1], %arg0 : !iree.list<f64>, f64 return %0 : !iree.list<f64> } ``` As part of doing this, I realized that it was time to formalize the IR form that we reach right before running TorchTo{Linalg,Std,...}. We now call it the "Torch backend contract". We then lower the "Torch backend contract" to the "npcomp backend contract", which involves the new TorchConversion (`torch_c`) dialect, which holds ops that need to operate on both the npcomp backend types (e.g. builtin tensors, i1, IREE list, etc.) and the `!torch` types. This made more sense, as I realized that if I didn't factor out `torch_c` then the Torch dialect would have a dependency on IREE dialect (we previously didn't notice this was an issue because we only depended on `builtin` types), which seemed wrong to me. Recommended review order: - TorchToIREE.cpp / `TorchToIREE/basic.mlir` - Look at the new structure of createTorchScriptToNpcompBackendPipeline. It now lives in TorchConversion/Transforms/Passes.cpp and cleanly calls into `Torch::createTorchScriptToTorchBackendPipeline` for the frontend lowering to the Torch backend contract. - Mechanical change extracting `torch_c.{to,from}_{i1,i64,f64,builtin_tensor,iree_list}` into a new TorchConversion dialect, and a few passes specific to the lowering from the Torch backend contract to the npcomp backend contract. - Minor fixes to TorchToLinalg.cpp to use unconverted operands (now that we convert lists as part of operand materialization, we need to use the original operands). Also added test for AtenMaxPool2dOp and fixed m_TorchConstantIntList. - TmpDeleteDeadIREELists pass. Temporary pass for deleting dead IREE lists that are created as part of operand materialization for conv/max pool/avg pool ops in TorchToLinalg.
2021-08-12 05:40:08 +08:00
%int1 = torch.constant.int 1
%int2 = torch.constant.int 2
%int3 = torch.constant.int 3
%int4 = torch.constant.int 4
%int5 = torch.constant.int 5
%int6 = torch.constant.int 6
%int7 = torch.constant.int 7
%int8 = torch.constant.int 8
%false = torch.constant.bool false
[MLIR][TORCH] Add decomposition of aten.adaptive_avg_pool2d op This commit adds the decomposition of `aten.adaptive_avg_pool2d` op into `aten.avg_pool2d` op. The current decomposition only supports cases where input size is equal to the output size. Signed-Off By: Vivek Khandelwal <vivek@nod-labs.com>
2022-05-13 20:06:24 +08:00
// CHECK: %[[C1:.*]] = torch_c.to_i64 %int1
// CHECK: %[[C2:.*]] = torch_c.to_i64 %int2
[Linalg] fix lowering reduce max with -inf (#2097)
2023-05-09 00:17:49 +08:00
// CHECK: %[[NEUTRAL:.*]] = arith.constant 0xFF800000 : f32
Bump LLVM to 881ff4e4ebe8cc0cc045c7c167cffb01f94f27f8 (#539)
2022-01-26 14:16:30 +08:00
// CHECK: %[[PADDED:.*]] = tensor.pad %{{.*}} low[0, 0, 5, 6] high[0, 0, 5, 6]
Bump LLVM at 8361c5da30588d3d4a48eae648f53be1feb5cfad
2022-03-16 18:44:23 +08:00
// CHECK: %[[OUT:.*]] = linalg.fill ins(%[[NEUTRAL]] : f32) outs(%{{.*}} : tensor<?x?x?x?xf32>) -> tensor<?x?x?x?xf32>
[MLIR][TORCH] Add decomposition of aten.adaptive_avg_pool2d op This commit adds the decomposition of `aten.adaptive_avg_pool2d` op into `aten.avg_pool2d` op. The current decomposition only supports cases where input size is equal to the output size. Signed-Off By: Vivek Khandelwal <vivek@nod-labs.com>
2022-05-13 20:06:24 +08:00
// CHECK: %[[T1:.*]] = arith.index_cast %[[C1]] : i64 to index
// CHECK: %[[T2:.*]] = arith.index_cast %[[C2]] : i64 to index
build: update llvm tag to 4546397e (#1502) This commit makes the following changes needed to update bump LLVM: - Replace `linalg.init_tensor` with `tensor.empty` (see: https://reviews.llvm.org/D135129) - Replace `NoSideEffect` with `Pure` (see https://reviews.llvm.org/D135505) - Replace `body` region accessor for `ReduceOp` and `ReduceWindowOp` with `getBody` - Fix incorrect use of `tosa::ReduceSumOp` in `AtenNativeLayerNormOp` conversion pattern. The result type of `tosa::ReduceSumOp` must have the same rank as the input type. (see: https://www.mlplatform.org/tosa/tosa_spec.html#_reduce_sum) Co-authored-by: Ashay Rane <ashay@users.noreply.github.com> Co-authored-by: Ashay Rane <ashay@users.noreply.github.com>
2022-10-18 12:22:53 +08:00
// CHECK: %[[INIT:.*]] = tensor.empty(%[[T1]], %[[T2]]) : tensor<?x?xf32>
Add TorchToIREE and factor out TorchConversion dialect. This converts a basic list op (torch.prim.ListConstruct) to the IREE dialect. ``` def forward(self, x: float): return [x, x] ``` turns into: ``` builtin.func @forward(%arg0: !torch.float) -> !torch.list<!torch.float> { %0 = torch.prim.ListConstruct %arg0, %arg0 : (!torch.float, !torch.float) -> !torch.list<!torch.float> return %0 : !torch.list<!torch.float> } ``` which turns into: ``` builtin.func @forward(%arg0: f64) -> !iree.list<f64> { %c1 = constant 1 : index %c0 = constant 0 : index %c2 = constant 2 : index %0 = iree.list.create %c2 : !iree.list<f64> iree.list.set %0[%c0], %arg0 : !iree.list<f64>, f64 iree.list.set %0[%c1], %arg0 : !iree.list<f64>, f64 return %0 : !iree.list<f64> } ``` As part of doing this, I realized that it was time to formalize the IR form that we reach right before running TorchTo{Linalg,Std,...}. We now call it the "Torch backend contract". We then lower the "Torch backend contract" to the "npcomp backend contract", which involves the new TorchConversion (`torch_c`) dialect, which holds ops that need to operate on both the npcomp backend types (e.g. builtin tensors, i1, IREE list, etc.) and the `!torch` types. This made more sense, as I realized that if I didn't factor out `torch_c` then the Torch dialect would have a dependency on IREE dialect (we previously didn't notice this was an issue because we only depended on `builtin` types), which seemed wrong to me. Recommended review order: - TorchToIREE.cpp / `TorchToIREE/basic.mlir` - Look at the new structure of createTorchScriptToNpcompBackendPipeline. It now lives in TorchConversion/Transforms/Passes.cpp and cleanly calls into `Torch::createTorchScriptToTorchBackendPipeline` for the frontend lowering to the Torch backend contract. - Mechanical change extracting `torch_c.{to,from}_{i1,i64,f64,builtin_tensor,iree_list}` into a new TorchConversion dialect, and a few passes specific to the lowering from the Torch backend contract to the npcomp backend contract. - Minor fixes to TorchToLinalg.cpp to use unconverted operands (now that we convert lists as part of operand materialization, we need to use the original operands). Also added test for AtenMaxPool2dOp and fixed m_TorchConstantIntList. - TmpDeleteDeadIREELists pass. Temporary pass for deleting dead IREE lists that are created as part of operand materialization for conv/max pool/avg pool ops in TorchToLinalg.
2021-08-12 05:40:08 +08:00
// CHECK: linalg.pooling_nchw_max {dilations = dense<[7, 8]> : vector<2xi64>, strides = dense<[3, 4]> : vector<2xi64>} ins(%[[PADDED]], %[[INIT]] : tensor<?x?x?x?xf32>, tensor<?x?xf32>) outs(%[[OUT]] : tensor<?x?x?x?xf32>) -> tensor<?x?x?x?xf32>
Elide `!torch.` prefix in nested dialect types. This leads to much more succinct types in many cases: ``` !torch.list<!torch.int> !torch.list<int> !torch.tuple<!torch.list<!torch.int>, !torch.list<!torch.int>> !torch.tuple<list<int>, list<int>> !torch.optional<!torch.list<!torch.int>> !torch.optional<list<int>> !torch.list<list<list<tensor>>> !torch.list<!torch.list<!torch.list<!torch.tensor>>> ``` I would like to take this further and allow omitting the `!torch.` prefix in all cases, but that's harder -- for example, we currently use `FuncOp` for functions, and so I don't think we can customize the printing there. It seems like it will be a longer road to getting that level of customization.
2022-03-16 07:22:56 +08:00
%kernel_size = torch.prim.ListConstruct %int1, %int2 : (!torch.int, !torch.int) -> !torch.list<int>
%stride = torch.prim.ListConstruct %int3, %int4 : (!torch.int, !torch.int) -> !torch.list<int>
%padding = torch.prim.ListConstruct %int5, %int6 : (!torch.int, !torch.int) -> !torch.list<int>
%dilation = torch.prim.ListConstruct %int7, %int8 : (!torch.int, !torch.int) -> !torch.list<int>
%4 = torch.aten.max_pool2d %arg0, %kernel_size, %stride, %padding, %dilation, %false : !torch.vtensor<[?,?,?,?],f32>, !torch.list<int>, !torch.list<int>, !torch.list<int>, !torch.list<int>, !torch.bool -> !torch.vtensor<[?,?,?,?],f32>
Add TorchToIREE and factor out TorchConversion dialect. This converts a basic list op (torch.prim.ListConstruct) to the IREE dialect. ``` def forward(self, x: float): return [x, x] ``` turns into: ``` builtin.func @forward(%arg0: !torch.float) -> !torch.list<!torch.float> { %0 = torch.prim.ListConstruct %arg0, %arg0 : (!torch.float, !torch.float) -> !torch.list<!torch.float> return %0 : !torch.list<!torch.float> } ``` which turns into: ``` builtin.func @forward(%arg0: f64) -> !iree.list<f64> { %c1 = constant 1 : index %c0 = constant 0 : index %c2 = constant 2 : index %0 = iree.list.create %c2 : !iree.list<f64> iree.list.set %0[%c0], %arg0 : !iree.list<f64>, f64 iree.list.set %0[%c1], %arg0 : !iree.list<f64>, f64 return %0 : !iree.list<f64> } ``` As part of doing this, I realized that it was time to formalize the IR form that we reach right before running TorchTo{Linalg,Std,...}. We now call it the "Torch backend contract". We then lower the "Torch backend contract" to the "npcomp backend contract", which involves the new TorchConversion (`torch_c`) dialect, which holds ops that need to operate on both the npcomp backend types (e.g. builtin tensors, i1, IREE list, etc.) and the `!torch` types. This made more sense, as I realized that if I didn't factor out `torch_c` then the Torch dialect would have a dependency on IREE dialect (we previously didn't notice this was an issue because we only depended on `builtin` types), which seemed wrong to me. Recommended review order: - TorchToIREE.cpp / `TorchToIREE/basic.mlir` - Look at the new structure of createTorchScriptToNpcompBackendPipeline. It now lives in TorchConversion/Transforms/Passes.cpp and cleanly calls into `Torch::createTorchScriptToTorchBackendPipeline` for the frontend lowering to the Torch backend contract. - Mechanical change extracting `torch_c.{to,from}_{i1,i64,f64,builtin_tensor,iree_list}` into a new TorchConversion dialect, and a few passes specific to the lowering from the Torch backend contract to the npcomp backend contract. - Minor fixes to TorchToLinalg.cpp to use unconverted operands (now that we convert lists as part of operand materialization, we need to use the original operands). Also added test for AtenMaxPool2dOp and fixed m_TorchConstantIntList. - TmpDeleteDeadIREELists pass. Temporary pass for deleting dead IREE lists that are created as part of operand materialization for conv/max pool/avg pool ops in TorchToLinalg.
2021-08-12 05:40:08 +08:00
return %4 : !torch.vtensor<[?,?,?,?],f32>
}