strisunshinewentingwang/pytorch

Pull Requests Commits

[Bug fix] sparse_mask: wrong intersection on CUDA (#94829)

nikitaved committed 1 year ago

3ace14eb

[FSDP] Fix `clip_grad_norm_()` when rank has no local gradients (#94835)

awgu committed 1 year ago

0c3ba785

[FSDP] Fix "use-after-free" in reshard logic (#94859)

awgu committed 1 year ago

8da776e3

Add further info to `masked_scatter` and `masked_scatter_` documention (#94545)

kiersten-stokes committed 1 year ago

5a545379

Update smoke test threshold (#94888)

weiwangmeta committed 1 year ago

5705199f

[ROCm] Pyt 2.0 rocm staging (#94660)

dllehr-amd committed 1 year ago

77d11355

[MPS] Block uint8 data type for unary and binary ops on macOS 12 (#94876)

DenisVieriu97 committed 1 year ago

71ec2617

Update ideep to add primitive cache for ARM (#94719)

yanbing-j committed 1 year ago

8261c600

Revert "Temporarily disable inductor torchbench test (#94873)"

pytorchmergebot committed 1 year ago

c10acb83

call `zero_grad` in foreach/fused optimizers tests (#94724)

crcrpar committed 1 year ago

e0a954f5

[ONNX] Fix assert in cat (#94870)

justinchuby committed 1 year ago

afadc369

Update vision commit pin (#94874)

malfet committed 1 year ago

3d5f4dcc

[CI] Install `pytorch-cuda` for conda testing (#94852)

malfet committed 1 year ago

117fafc2

Temporarily disable inductor torchbench test (#94873)

kit1980 committed 1 year ago

79b7c697

Make _simplified kwarg private (#94782)

ezyang committed 1 year ago

abf59f57

PT2/TorchScript interoperability fix (#94678)

jansel committed 1 year ago

ae57bd66

[ONNX] Wrap op validation inputs and add export_options.py and function_dispatcher.py (#94721)

titaiwangms committed 1 year ago

b6443fca

sym_int simplification for integer args, attempt 3 (#94799)

ngimel committed 1 year ago

5bc72bd0

[inductor] Disable developer warnings for "2.0.0" version (#94845)

jansel committed 1 year ago

65b99832

add reproducibility notes to nn.UnpoolND operations (#94629)

drisspg committed 1 year ago

7f7f91e3

Fix layout/device checks in sparse-dense addmm (#94843)

amjames committed 1 year ago

7c44823a

Switch Docker release to CUDA 11.7 (#94818)

atalman committed 1 year ago

40cb494b

[ROCm] hipGraph support for pytorch mainline (#88202)

dllehr-amd committed 1 year ago

98012e4a

[torchgen] Loosen the restriction for only allowing 2 nested namespaces for kernels (#94834)

larryliu0820 committed 1 year ago

79783a51

Preloads more nvidia pypi library for multi arch distributions (#94355)

syed-ahmed committed 1 year ago

7ef76ce6

Convert operator.not_ to torch.logical_not (#94626)

angelayi committed 1 year ago

97510c6d

[ROCm] Added MIOpen header files to installation package for ROCm. (#92969)

wenchenvincent committed 1 year ago

69bcefce

Use s3 for some test infra files (#94642)

clee2000 committed 1 year ago

98929980

add backwards for gelu and relu on nested tensors. (#94776)

drisspg committed 1 year ago

63bf7674

Improve leaky relu doc (#94090)

albanD committed 1 year ago

b7e1477e

Older