This repository has been archived by the owner on Jan 30, 2025. It is now read-only.
Lower torch transposed convolution to a custom TCP op #25
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
srinathava
commented
Jan 11, 2024
Comment on lines
+166
to
+169
| auto addOperand = [&](std::string name, Value value) { | ||
| operandNames.push_back(name); | ||
| operands.push_back(value); | ||
| }; |
There was a problem hiding this comment.
Non-blocking and could be addressed in a follow-on PR:
This would be nice to factor out as a utility, so they can be reused by other TcpCustomOp conversion patterns.
Comment on lines
182
to
190
| auto addListOfIntAttr = [&](const std::string &name, Value value) { | ||
| SmallVector<int64_t> valueInt; | ||
| if (!matchPattern(adaptor.getStride(), | ||
| m_TorchListOfConstantInts(valueInt))) | ||
| return rewriter.notifyMatchFailure(op, std::string("non-const") + name + | ||
| "list unsupported"); | ||
| attrs.push_back( | ||
| rewriter.getNamedAttr(name, rewriter.getIndexArrayAttr(valueInt))); | ||
| return success(); | ||
| }; |
Comment on lines
77
to
81
| %input2 = torch.vtensor.literal(dense<0.0> : tensor<1x9x16x1600xf32>) : !torch.vtensor<[1,9,16,1600],f32> | ||
| %weights2 = torch.vtensor.literal(dense<0.0> : tensor<32x9x3x3xf32>) : !torch.vtensor<[32,9,3,3],f32> | ||
| %int0x0 = torch.prim.ListConstruct %int0, %int0 : (!torch.int, !torch.int) -> !torch.list<int> | ||
| %none = torch.constant.none | ||
| %output2 = torch.aten.convolution %input2, %weights2, %none, %int1x1, %int1x1, %int1x1, %false, %int0x0, %int1 : !torch.vtensor<[1,9,16,1600],f32>, !torch.vtensor<[32,9,3,3],f32>, !torch.none, !torch.list<int>, !torch.list<int>, !torch.list<int>, !torch.bool, !torch.list<int>, !torch.int -> !torch.vtensor<[1,32,16,1600],f32> |
There was a problem hiding this comment.
Please consider splitting this into a separate test (in the same file) for readability and/or ease of maintenance. You may use self-explanatory test names, like
func.func @torch.aten.regular_convolution
...
func.func @torch.aten.transposed_convolution
...
There was a problem hiding this comment.
good suggestion. I split it and also split the CHECK line into multiple lines to make it a bit clearer.
|
|
||
| // --- | ||
|
|
||
| // CHECK: tcp.custom_op("torch.aten.convolution") %{{.*}}, %{{.*}}, %{{.*}} {dilation = [2 : index, 2 : index], groups = 1 : i64, output_padding = [2 : index, 2 : index], padding = [2 : index, 2 : index], stride = [2 : index, 2 : index], torch_operand_names = ["input", "weight", "bias"], transposed = true} : tensor<1x64x1x100xf32>, tensor<64x64x3x3xf32>, tensor<64xf32> -> tensor<1x64x2x200xf32> |
There was a problem hiding this comment.
I was wondering how the test passes despite the bug above. It seems to be because padding is being populated from stride and the check statement propagates the bug.
// This should've been padding = [1 : index, 1 : index]
padding = [2 : index, 2 : index]
There was a problem hiding this comment.
:( This PR is turning out to be a bit of typo central. Thanks for the careful review.
| @@ -19,6 +19,7 @@ | |||
| #include "torch-mlir/Dialect/Torch/Utils/Utils.h" | |||
|
|
|||
| #include "llvm/ADT/StringSet.h" | |||
| #include "llvm/Support/Debug.h" | |||
There was a problem hiding this comment.
Just checking, is this still needed?
sjain-stanford
approved these changes
Jan 12, 2024
srinathava
added a commit
to srinathava/mlir-tcp
that referenced
this pull request
Jun 24, 2024
…ion#25) As titled, lower torch transposed convolution to a custom TCP op to avoid a mis-compilation in `TorchToTosa`. --------- Co-authored-by: Srinath Avadhanula <srinath.avadhanula@getcruise.com>
srinathava
pushed a commit
to srinathava/mlir-tcp
that referenced
this pull request
Jun 24, 2024
…ruise-automation#25) (cruise-automation#10) As titled, lower torch transposed convolution to a custom TCP op to avoid a mis-compilation in `TorchToTosa`. Cherry-pick from upstream: cruise-automation#25 --------- Co-authored-by: Srinath Avadhanula <srinath.avadhanula@getcruise.com>
srinathava
pushed a commit
to srinathava/mlir-tcp
that referenced
this pull request
Jun 24, 2024
Remove `LRDPrefilterPrediction` related content Test from c/c: `bazel test @mlir-tcp//test/...`
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
As titled, lower torch transposed convolution to a custom TCP op to avoid a mis-compilation in
TorchToTosa.