strisunshinewentingwang/pytorch

Pull Requests Commits

[ATen][AMD] revert a change of USE_DIRECT_NVRTC (#74194)

shintaro-iwasaki committed 2 years ago

102add66

Preserve codegen on fx graph in transformer (#74189)

842974287 committed 2 years ago

f98b316f

Dispatch from torch.Tensor.to_sparse_coo to to_sparse

cpuhrsch committed 2 years ago

4de9cb9a

[PyTorch] Add unit test for c10::Synchronized<T> (#74110)

dhruvbird committed 2 years ago

606c26d3

[PyTorch] Use c10::Synchronized<T> in RWSafeLeftRightWrapper (#74109)

dhruvbird committed 2 years ago

977cc480

[PyTorch] Update Synchronized<T>::withLock() to return the type/value from the aceepted callable (#74108)

dhruvbird committed 2 years ago

e1eb876a

[PyTorch] [Model Tracer] Use c10::Synchronized<T> for build features tracer (#74107)

dhruvbird committed 2 years ago

44ef979d

[PyTorch] [Model Tracer] Use c10::Synchronized<T> for custom classes tracer (#74106)

dhruvbird committed 2 years ago

3674f92d

[PyTorch] [Model Tracer] Use c10::Synchronized<T> for kernel dtype tracer (#74105)

dhruvbird committed 2 years ago

cf461a8b

Use assertEqual consistently in test_sparse_csr.py

cpuhrsch committed 2 years ago

a346a181

Excluding ASAN and periodic jobs from slow job calculation (#74253)

atalman committed 2 years ago

2e15e16f

Support running pipelines on main in .jenkins and tools

musebc committed 2 years ago

0effac2b

[Quant][core][refactorization] Refactored qconv_unpack.cpp into an implementation file and higher level call registration and definition file (Reland #73773) (#74227)

dzdang committed 2 years ago

ce7041d2

[Quant] qconv: fix xnnpack operator caching (#74217)

digantdesai committed 2 years ago

f8c1acea

Add operator selection ability to gen_unboxing (#74271)

priyaramani committed 2 years ago

92751498

Improve error message of loading saved TS module out of support window (#74228)

gmagogsfm committed 2 years ago

d5744f47

[Quant][core][gpu][improvement] Refactored implementation for conv2d_cudnn to use packed parameters (Reland PR#73510) (#74220)

dzdang committed 2 years ago

a2f701d4

[PyTorch GPU Allocator] Better use of blocks with rounding of allocation sizes (#74213)

banitag1 committed 2 years ago

ac3effd1

[ao] Fixing obs insertion through dtype propagation (#73274)

HDCharles committed 2 years ago

c1d070d0

Automated submodule update: FBGEMM (#74104)

facebook-github-bot committed 2 years ago

3c837a8b

[GHA][BE] Delete unused definitions (#74270)

malfet committed 2 years ago

Verified d94f440c

[quant][fx] Allow incrementally remove the items in quantization_patterns.py (#74210)

jerryzh168 committed 2 years ago

ca4348f6

[quant] Fix implementation for `output_quantized_idxs` in convert (#74140) (#74229)

jerryzh168 committed 2 years ago

9a0b7b47

Revert "Add facebook-github-bot to superuser approver list"

pytorchmergebot committed 2 years ago

e109d66f

[caffe2] Fix alias analysis for quantization compression ops (#74169)

Mike Iovine committed 2 years ago

6bd4376c

release: Add convenience script for branch cutting

seemethere committed 2 years ago

ddb34e7b

Add facebook-github-bot to superuser approver list

bigfootjon committed 2 years ago

6b92abe0

Revert D34856571: [pytorch][PR] Replace `get_all_` type macros with the ATen dispatch macros.

malfet committed 2 years ago

ef066f08

[SR] Avoid allocating rstd/mean in layer_norm (#73606)

Mike Iovine committed 2 years ago

f14a0be3

Performance and memory improvements to batched torch.linalg.solve (2nd attempt) (#71756)

IvanYashchuk committed 2 years ago

064206df

Older