strisunshinewentingwang/pytorch

Pull Requests Commits

avoiding adding some functions to the public python API before 1.11 release (#72543)

bdhirsh committed 2 years ago

f87f753b

[ROCm] Enable sort operator BF16 support (#71226)

pruthvistony committed 2 years ago

aa44480b

[Quant][core][devs] Separated implementations for quantized & non-quantized tensors in index_select_cuda (#72407)

dzdang committed 2 years ago

9981aade

Properly initialize `grad_weight` in `raw_cudnn_convolution_backward_weight_out` (#72157)

eqy committed 2 years ago

e7985e3c

Fix type annotation for `torch.backends.cudnn.allow_tf32` (#72757)

lkct committed 2 years ago

7d542a4f

[ROCM] Navi21 Enablement 2: Depthwise kernels (#72682)

micmelesse committed 2 years ago

89c934f4

DispatchKeySet perf improvements (#72403)

bdhirsh committed 2 years ago

8aa3620d

free up dispatch key space (in C++) (#72402)

bdhirsh committed 2 years ago

66902560

Implement Tanh Gelu Approximation (#61439)

rdspring1 committed 2 years ago

4f8b986e

Mobile upgrader: add GELU test modules

jbschlosser committed 2 years ago

64faa043

Update from_dlpack tests and documentation (#70543)

kurtamohler committed 2 years ago

47c69933

Automated submodule update: FBGEMM (#72769)

facebook-github-bot committed 2 years ago

1776caf3

Back out "[SR] Make fused_sigrid_transforms work on graph outputs" (#72767)

Mike Iovine committed 2 years ago

c9ccb850

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

CodemodService FBSourceClangFormatLinterBot committed 2 years ago

5a95ffbd

[FSDP] Use unflatten_parameter in _summon_full_parameters (#72467)

fegin committed 2 years ago

c73cc92e

Automated submodule update: FBGEMM (#72764)

facebook-github-bot committed 2 years ago

81a1330b

Allow forking until a worker thread is created in autograd engine (#72689)

albanD committed 2 years ago

4d04ef62

Clean up use of cpu ready queue in autograd engine (#72688)

albanD committed 2 years ago

a8774414

Remove un-used function in autograd engine (#72687)

albanD committed 2 years ago

13125247

Use default value on empty llvm_code_path (#72758)

priyaramani committed 2 years ago

444191de

Deprecate torch.jit.quantized (#72690)

peterbell10 committed 2 years ago

dce84b9c

Add autograd tests for addmm, addmv, mm, mv and CSR matrix input (#71949)

IvanYashchuk committed 2 years ago

fb7c4780

[caffe2] Allow LpNorm to accept empty tensor (#72660)

Dmytro Dzhulgakov committed 2 years ago

6b24d7e4

[GH1] Ignore "COMMENTED" reviews

malfet committed 2 years ago

ef92b549

[GH1] Enable cross-repo merges

malfet committed 2 years ago

0c95209f

[Doc] Better formatting in autograd.rst (#72586)

felixdivo committed 2 years ago

340fae43

Release documentation update to include latest GHA changes (#72740)

atalman committed 2 years ago

11359e14

[ROCm] Enable skipped ROCm unit tests (#67706)

dllehr-amd committed 2 years ago

28549b61

[ONNX] Minor doc update (#69501) (#69550)

BowenBao committed 2 years ago

9257de7e

[ONNX] Add Concat to Scalar type analysis JIT pass (#69227) (#69548)

BowenBao committed 2 years ago

7884c2bb

Newer Older