strisunshinewentingwang/pytorch

Pull Requests Commits

Avoid device casting for all singleton tensors in optimizer states (#91454)

jbschlosser committed 1 year ago

1e725c97

[MPS] Fix the crash in max_out() caused by cached key conflict (#91520)

razarmehr committed 1 year ago

97925506

Fix typo in _lobpcg.py (#91641)

eltociear committed 1 year ago

ce9963e6

Adds more nvidia pypi dependencies (#89944)

syed-ahmed committed 1 year ago

66b33253

squeeze: allow squeezing multiple dimensions at once (#89017)

peterbell10 committed 1 year ago

e26cb066

Vectorize norm(double, p=2) on cpu (#91502)

lezcano committed 1 year ago

3120054c

Remove python ddp (#91663)

zhaojuanmao committed 1 year ago

2004df90

quant: make various configs printable (#91419)

vkuzo committed 1 year ago

ebb7f20a

Run jit legacy tests sequentially (#91518)

huydhn committed 1 year ago

316ba9e6

[MPS] Register norm_dtype_out_mps and cdist (#91643)

DenisVieriu97 committed 1 year ago

80394bb7

Make torch.device usable as a context manager (#91525)

ezyang committed 1 year ago

619d52a5

[Inductor] add missing ops for cpp vectorization overrides (#90750)

Valentine233 committed 1 year ago

aa0ca994

inductor: add conv+hardsigmoid fusion for cpu path (#91433)

XiaobingSuper committed 1 year ago

1d2bfea3

Revert "Populate the eviction_policy field for load/store properly (#91316)"

pytorchmergebot committed 1 year ago

6f9a4ae5

Add an env variable to disable addmm_cuda_lt kernel (#91436)

zhaojuanmao committed 1 year ago

e116f1a3

[functorch] add new ensembling api, demonstrate in example (#88850)

samdow committed 1 year ago

162474d7

[functorch] add functorch functional_call, update tests to test this (#89213)

samdow committed 1 year ago

c5e5916f

[autograd.Function] Add docs on the functorch interaction (#91452)

zou3519 committed 1 year ago

264f5ed5

Remove CircleCI ios PR jobs (#91638)

clee2000 committed 1 year ago

31a69993

[MPS] Add nonzero mps support (#91616)

DenisVieriu97 committed 1 year ago

38de981e

[cuBLAS] (re-open) Fix default cuBLAS workspace size and parsing for multiple workspaces (#91564)

eqy committed 1 year ago

97ff20d7

Revert "Avoid copies in matmul (#76828)"

pytorchmergebot committed 1 year ago

0a6053e9

[inductor] Check for BackendCompilerFailed on CI (#91634)

desertfire committed 1 year ago

6bf0e3b6

implement ordering (#91362)

drisspg committed 1 year ago

3a60debe

refactor show_traces in memory_tracker (#90145)

zhaojuanmao committed 1 year ago

743c3855

Revert "Dispatch the auxiliary frobenius_norm and nuclear_norm to better implementations and deprecate them (#81763)"

pytorchmergebot committed 1 year ago

b6bb726c

[Inductor] Move graph.lint() in Intel's FX Passes to the End of Loop to Reduce Compile Time (#91179)

jiawenliu64 committed 1 year ago

57b7f33b

disabled flaky c2 test (#91640)

ngimel committed 1 year ago

818079dc

CSC/BSC -> COO coalesce fix (#91440)

nikitaved committed 1 year ago

7ef7c57a

Revert D42051833: Multisect successfully blamed D42051833 for test or build failures (#91458)

weiwangmeta committed 1 year ago

47095237

Newer Older