strisunshinewentingwang/pytorch

Pull Requests Commits

[7/N] [Dispatchable Collectives] Update reduce with CPU / CUDA implementations (#83916)

H-Huang committed 1 year ago

8a1fc5d2

Reland 2 of Merge more symbolic meta kernels and symint changes from branch (#86334) (#86488)

albanD committed 1 year ago

978b46d7

Reland 3 of Symintify getitem and add the required helper functions (#86207) (#86487)

albanD committed 1 year ago

55663b7f

fix some composite compliance ops for functionalization (#86470)

bdhirsh committed 1 year ago

4a5fdc56

[FSDP][1/N] Retire `FlattenParamsWrapper` (#86117)

awgu committed 1 year ago

5102f0cf

[xla hash update] update the pinned xla hash (#86099)

pytorchmergebot committed 1 year ago

bf7c46fa

[FSDP] Add `low_prec` prefix to param and reduce dtype varnames (#86512)

awgu committed 1 year ago

5844f00b

[FSDP] Remove `utils.py` (moved to `_utils.py`) (#86528)

awgu committed 1 year ago

cc5de7f1

torchdynamo: add linear eltwise fusion kernel (#85622)

chunyuan-w committed 1 year ago

c6b7c338

[torchdynamo hash update] update the pinned torchdynamo hash (#86567)

pytorchmergebot committed 1 year ago

ec2d22ec

BlasKernel: Improve gemm's inner dot product when a is transposed (#80977)

peterbell10 committed 1 year ago

753536b7

mkl: Use per-operator headers (#75570)

peterbell10 committed 1 year ago

a45fead6

symintify unbind_backward and tensor_split (#86357)

anjali411 committed 1 year ago

c89d286a

Add __all__ to torch.{autograd, fx, cuda} submodules (#85343)

anjali411 committed 1 year ago

a6c0442c

[BE] Remove remaining cuda-11.3 builds (#86540)

malfet committed 1 year ago

6aec0d3d

Quantized: Use per-operator headers (#75569)

peterbell10 committed 1 year ago

7134b9bc

[MPS] Fix printTensor() for MPS (#86534)

malfet committed 1 year ago

67434c70

[vision hash update] update the pinned vision hash (#86490)

pytorchmergebot committed 1 year ago

9998f910

[torchdynamo hash update] update the pinned torchdynamo hash (#86489)

pytorchmergebot committed 1 year ago

92ac84c9

QuantizedCPU: Use per-operator headers (#71217)

peterbell10 committed 1 year ago

492d1be5

cuDNN/miopen: Use per-operator headers (#71216)

peterbell10 committed 1 year ago

4bfe2a24

Re-land*4 "SymIntify cat and narrow" (#86468)

ezyang committed 1 year ago

33f0e98a

Revert "Re-enable torchdynamo tests (#86297)"

pytorchmergebot committed 1 year ago

8ea2ed0f

Enable aten-aten decomps (#85921)

eellison committed 1 year ago

d3f7c34c

[FSDP] Add `keep_low_precision_grads` support when CPU offloading (#86495)

awgu committed 1 year ago

af9c6bc8

Revert "Enable aten-aten decomps (#85921)"

pytorchmergebot committed 1 year ago

7ec12a55

[vulkan] Add buffer to buffer copies (#86424)

SS-JIA committed 1 year ago

b0ceb8ea

[vulkan] Clean up convolution code (#86423)

SS-JIA committed 1 year ago

511d81cd

make g2p ~30% faster on mobile by suppressing a log (#85907)

codyohl committed 1 year ago

b645c237

[JIT] Allow freezing modules that contain mutable interfaces (#86039)

davidberard98 committed 1 year ago

bac26155

Newer Older