strisunshinewentingwang/pytorch

Pull Requests Commits

[optim][adam] use fastest impl whenever possible, add util

janeyx99 committed 1 year ago

acb59584

Revert "Add cudnn install 8.7.0.84 for CUDA 11.8 (#93086)"

pytorchmergebot committed 1 year ago

62aa4e09

Repair the path to jni.h for libtorch windows build (#93057)

cristianPanaite committed 1 year ago

d3049378

Explicit Name needed to run with buck test (#93035)

mikekgfb committed 1 year ago

64d0624c

Add cudnn install 8.7.0.84 for CUDA 11.8 (#93086)

atalman committed 1 year ago

3a10bf79

[fix] nn c++ : segfault in modulelist and moduledict (#93074)

kshitij12345 committed 1 year ago

68a98537

Improve autograd doc on complex numbers (#93065)

felixdivo committed 1 year ago

219e9533

Enable Kineto in OSS builds by fixing build condition (resubmit) (#93033)

dzhulgakov committed 1 year ago

5105a8d3

[inductor] Clean up TRITON_CACHE_DIR (#92879)

xw285cornell committed 1 year ago

070163fb

[FX][Quant] Enable FX quant for patterns like x.view(x.size(...), ...) (#90001)

Xia-Weiwen committed 1 year ago

6fa84fde

[FSDP][optim_state_dict] Ensure correct devices for tensors when doing all_gather (#92992)

fegin committed 1 year ago

a4238976

[FSDP][optim_state_dict] Use all_gather to deal with uneven size tensors (#92991)

fegin committed 1 year ago

8b1b47c3

More tidy fixes (#93069)

cyyever committed 1 year ago

f172feae

Don't graph break on patched module methods (#93115)

williamwen42 committed 1 year ago

5bae5805

[ FL-gradient quantization] Adding QNN unpack feature (#92714)

kwanghoon-meta committed 1 year ago

a2e0f8e5

Fix BC-breaking change introduced by #91499 (#93091)

malfet committed 1 year ago

661800a2

fixing flag to skip nvfuser_tests build (#93080)

jjsjann123 committed 1 year ago

7fade4f7

[vision hash update] update the pinned vision hash (#93114)

pytorchmergebot committed 1 year ago

e2739372

Install Torchvision in all Linux shards (#93108)

huydhn committed 1 year ago

074f5ce0

Get rid of dedicated inductor dynamic_shapes config (#93076)

ezyang committed 1 year ago

025ef99d

[dtensor][7/N] remove backend in with_comms (#93040)

wanchaol committed 1 year ago

f3fcc806

[DTensor] implement dist_cat as a sharding prop rule (#92677)

XilunWu committed 1 year ago

8b3e01cd

[ONNX] Export 'aten::index_put(self, mask, v)' when rank(mask) < rank(self) (#92862)

BowenBao committed 1 year ago

24172eeb

Add kwargs support to torch.export() API (#92013)

Thiago Crepaldi committed 1 year ago

95dfad9d

[ci] Move sm86 from trunk to pull (#93085)

clee2000 committed 1 year ago

ae171cf6

Fix topk IMA (#93095)

zasdfgbnm committed 1 year ago

d1807dc1

Make __deepcopy__ of GraphModule able to handle circular reference. (#93038)

qihqi committed 1 year ago

8d7f9e2f

[CI] Move parallel native builds to 3.8 (#93103)

malfet committed 1 year ago

ceb44350

[Reland] aot autograd explicitly errors on double backward (#92893)

soulitzer committed 1 year ago

f6f46ba3

Revert "Disable torch_jit_fuser_te for dynamo CI (#92945)"

pytorchmergebot committed 1 year ago

913cf290

Older