strisunshinewentingwang/pytorch

Pull Requests Commits

Keep alive graph when creating iterators from it (#51951)

t-vi committed 3 years ago

bd624810

Fix test time history report if no ancestor report (#52054)

samestep committed 3 years ago

ce8ba5f3

Silence harmless error logs of TensorPipe agent during shutdown (#51785)

lw committed 3 years ago

a1c67b07

Avoid TensorPipe agent spamming logs when unable to guess IP address (#51784)

lw committed 3 years ago

b7b944a3

Use CUDA 11.2 for nightly docker build. (#51990)

xuzhao9 committed 3 years ago

03e82f79

[torch script] Add pure list producing ops to alias analysis (#51999)

bwasti committed 3 years ago

c4a8f0ce

Add benchmark for torch.nn.functional.interpolate

NicolasHug committed 3 years ago

50e6f0fd

Early terminate CUDA on common_utils TestCases (#50914)

Rong Rong (AI Infra) committed 3 years ago

c1b7ca80

OpInfo: Added clamp and trunc tests with aliases (#51167)

vfdev-5 committed 3 years ago

8b0cb5ed

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

generatedunixname89002005325676 committed 3 years ago

3cf78395

Warn about floor_divide performing incorrect rounding (#50281) (#50281)

Mike Ruberry committed 3 years ago

594a66d7

Adding support for comparing two bool varibales (#51844)

nikithamalgi committed 3 years ago

9c0caf03

[te] Benchmark vml-based logit (#51771)

bertmaher committed 3 years ago

602434bc

[te] Implement log approximation using the VML approach (#51752)

bertmaher committed 3 years ago

2e35fe95

[te] Introduce likely/unlikely CompareSelect hint (#51751)

bertmaher committed 3 years ago

ff73be7e

[te][llvm] Generate arithmetic vs logical right shift as appropriate (#51749)

bertmaher committed 3 years ago

74082f0d

Back out "Revert D26009829: Optimize relu on cpu using clamp_min" (#51819)

bertmaher committed 3 years ago

0620c96f

fake_quant cachemask: remove Python bindings (#51878)

vkuzo committed 3 years ago

33afb5f1

[model loading] Add max_batch_size override for batch size exploration

ChunliF committed 3 years ago

5f9fb93c

Migrate masked_fill__cuda to ATen (#51404)

kshitij12345 committed 3 years ago

76866291

ns_eager: rename Logger I/O var names to logger_cls (#51359)

vkuzo committed 3 years ago

929b91a2

Automated submodule update: FBGEMM (#52014)

facebook-github-bot committed 3 years ago

5a9bac58

add more logging fields that can be set in construction time (#51260)

zhaojuanmao committed 3 years ago

18e0a613

[FX] Generalize dict key check in `create-arg` (#51927)

James Reed committed 3 years ago

d23cb940

[FX][EZ] Fix tuple type annotations (#52010)

James Reed committed 3 years ago

256f93fb

[FX] Fix leaf modules in Transformer (#51998)

James Reed committed 3 years ago

d4e84b0c

[PyTorch] Re-order TensorImpl fields to save a word (#50920)

swolchok committed 3 years ago

d5a9627f

[FX] Make some modifications to limitation section (#51928)

Chillee committed 3 years ago

475278f1

Let child CUDAFuture wait for parent CUDAFuture's CUDAEvents (#51820)

mrshenli committed 3 years ago

3af7b673

[ROCm] add 4.0.1 docker image (#51507)

jeffdaily committed 3 years ago

c6b4fc8a

Newer Older