strisunshinewentingwang/pytorch

Pull Requests Commits

[quant][be] Add a check in prepare_qat to make sure the model is in training mode (#69879)

jerryzh168 committed 2 years ago

b7259b86

Add conversion of torch.permute to acc_ops.permute (#70294)

dborkovic committed 2 years ago

2806d821

fix csv output (#70291)

Krovatkin committed 2 years ago

Verified 1e5bfdd9

make inverse call linalg_inv (#70276)

ngimel committed 2 years ago

56969bf8

[nn] TransformerEncoderLayer: no-batch-dim (#69291)

kshitij12345 committed 2 years ago

4db3a8fc

Remove unused CUDASolver.h from SparseCUDABlas (#70281)

malfet committed 2 years ago

69b37a16

Install TensorRT lib on oss docker and enable fx2trt unit test (#70203)

wushirong committed 2 years ago

31c7e5d6

Codegen constant_pad_nd

desertfire committed 2 years ago

9f8eaa77

[AutoAccept][Codemod][FBSourceClangFormatLinter] Daily `arc lint --take CLANGFORMAT`

CodemodService FBSourceClangFormatLinterBot committed 2 years ago

b5f71375

Fix some Composite Compliance problems with binary_cross_entropy backward (#70198)

zou3519 committed 2 years ago

29f1ccc8

[DataPipe] removing unbatch_level from .groupby (#70249)

NivekT committed 2 years ago

75dbe88b

[LTC] Upstream LTCTensorImpl (#70062)

Jiewen Tan committed 2 years ago

e02d836c

[StaticRuntime] Add out-variant support for TensorExprDynamicGroup op (#69479)

navahgar committed 2 years ago

633f770c

[jit] Handle output tensor being passed in as inputs to TensorExprDynamicGroup (#69478)

navahgar committed 2 years ago

7d4db93a

[nnc] Add a run method to TensorExprKernel that takes in output tensors (#69477)

navahgar committed 2 years ago

4dec15e6

[jit] Add a new op that composes all of the dynamic shape logic (#69476)

navahgar committed 2 years ago

0bdf4702

Back out "[Quant] Added 4 bit support for embedding quantized module" (#70273)

digantdesai committed 2 years ago

b613fbdb

Back out "[Quant][Eager] Added 4 bit support for eager mode quantization flow" (#70272)

digantdesai committed 2 years ago

47ba28f3

Back out "[Quant][fx] Added test for quint4x2 for fx graph mode quantization" (#70274)

digantdesai committed 2 years ago

a86f9806

Revert D33246843: [pytorch][PR] Implementation of Wishart distribution

Arvind Kannan committed 2 years ago

6217fee9

[GHA] Fix doc push jobs (#70269)

malfet committed 2 years ago

2d509ff3

[Operator Versioning][Edge] Reorganize upgrader initialization logic for thread safety (#70225)

cccclai committed 2 years ago

591ca4d6

Extend autograd functional benchmarking to run vectorized tasks (#67045)

soulitzer committed 2 years ago

21c6de9f

[shard] fix named_params_with_sharded_tensor (#70228)

wanchaol committed 2 years ago

82c5f298

[DataPipe] adding a finally statement to ensure hook is reset (#70214)

NivekT committed 2 years ago

74c834e0

Fixed typo in torch check for cdist (#70178)

vfdev-5 committed 2 years ago

23902fb8

Implementation of Wishart distribution (#68588)

nonconvexopt committed 2 years ago

a217a62e

[reland] Support torch.equal for ShardedTensor. (#70145)

pritamdamania committed 2 years ago

0544f975

[Operator Versioning] Split the upgrader test to a separate file and cover mobile part (#70090)

cccclai committed 2 years ago

c321d4c1

[StaticRuntime] Add TensorExpr fusion with dynamic shapes in SR (#69475)

navahgar committed 2 years ago

a6f95315

Newer Older