13 Pull requests merged by 10 people

• [test] add unified jit importer test config

  #4083 merged Mar 10, 2025

• Add Windows build workflow in the CI.

  #4041 merged Mar 9, 2025

• Add support for transposed grouped convolution in torch to linalg lowering

  #4056 merged Mar 6, 2025

• [docs] Add "Enable Quicker Builds" section to development.md

  #4074 merged Mar 5, 2025

• Integrate LLVM at 6d847b1aada50d59c3e29f2e7eff779c0ee8182c

  #4073 merged Mar 5, 2025

• Extend to 1D and 3D the torch to linalg lowering of the average pool operator with count_include_pad = false

  #4035 merged Mar 5, 2025

• APInt fix

  #4070 merged Mar 4, 2025

• Bump the github-actions group with 5 updates

  #4065 merged Mar 4, 2025

• build: manually update PyTorch version

  #4069 merged Mar 4, 2025

• [torch] Rework torch.repeat to not broadcast unary case

  #4061 merged Mar 4, 2025

• [docs] Restructure "Set up pre-commit hooks" section in development.md

  #4068 merged Mar 3, 2025

• Add runtime asserts for aten.reflection.pad_2d

  #4057 merged Mar 3, 2025

• Update shape_inference.cpp for randn.generator

  #4062 merged Mar 3, 2025

10 Pull requests opened by 10 people

• [ONNX] Handle dynamic shaped inputs for conv

  #4066 opened Mar 3, 2025

• Add atleast3d op

  #4067 opened Mar 3, 2025

• [TorchToLinalg] Support non-zero rank `NumToTensor.Scalar`

  #4071 opened Mar 3, 2025

• [TOSA] TOSA updates for LLVM hash 6d847b1

  #4072 opened Mar 3, 2025

• Use all frames of the stack trace when importing

  #4075 opened Mar 4, 2025

• Add shape inference for AtenAsStrided

  #4076 opened Mar 5, 2025

• Add support for Quantization/Dequantization

  #4078 opened Mar 6, 2025

• [ONNX] Add per-column quantization support for Onnx.QLinearMatMul op

  #4080 opened Mar 7, 2025

• [docs] Move extra cmake command options in development.md

  #4081 opened Mar 7, 2025

• Initial commit does not build

  #4082 opened Mar 7, 2025

2 Issues opened by 2 people

• testcase failing due to commit 7b23a1f5d87064daded4b89487dc968fd933051e

  #4079 opened Mar 7, 2025

• Error when running build_tools/update_torch_ods.sh

  #4077 opened Mar 6, 2025

9 Unresolved conversations

Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.

• [ONNX] Add Onnx->Torch lowering for GroupQueryAttention op

  #4006 commented on Mar 7, 2025 • 8 new comments

• [ONNX] Add per channel quantization support for Onnx.QLinearConv op

  #3917 commented on Mar 9, 2025 • 4 new comments

• Use op.dtype to create aten.empty.memory_format during decomposition.

  #3941 commented on Mar 4, 2025 • 2 new comments

• [MLIR][TORCH] Add support for none result for convolution_backward op

  #4063 commented on Mar 9, 2025 • 1 new comment

• torch.cond operator not supported on a simple example

  #4028 commented on Mar 4, 2025 • 0 new comments

• Compile torch.nn.functional.scaled_dot_product_attention failed

  #3473 commented on Mar 8, 2025 • 0 new comments

• [docs] Refreshes development.md

  #3932 commented on Mar 7, 2025 • 0 new comments

• support decomposition of aten.adaptive_max_pool2d

  #3954 commented on Mar 10, 2025 • 0 new comments

• [onnx_importer] Fix constant bool tensor importing

  #4025 commented on Mar 4, 2025 • 0 new comments