strisunshinewentingwang/pytorch

Pull Requests Commits

Add torch::cuda::ncll::all2all (#45900)

malfet committed 3 years ago

c19b9cd1

Add `tensor_split` function, based on `numpy.array_split` (#45168)

kurtamohler committed 3 years ago

ef4817fe

Workaround for cublas bug for 45724 (#46001)

zasdfgbnm committed 3 years ago

b2bff9e4

codegen: Improve array default handing (#45163)

peterbell10 committed 3 years ago

8d14b50e

[FX] Preserve type annotations on generated code in Graph (#45880)

James Reed committed 3 years ago

00b8ebe6

Add `[zc]heevd` to the list of MKL symbols exported from torch_cpu (#46002)

malfet committed 3 years ago

81d40aaf

Fix cholesky TF32 tests (#45492)

invalid-email-address committed 3 years ago

c59c4b0d

CUDA BFloat16 support of clamp, remainder, lshift, rshift (#45247)

zasdfgbnm committed 3 years ago

903acc6b

Fix distributed documentation for asynchronous collective Work objects (#45709)

rohan-varma committed 3 years ago

154347d8

[NNC] Registerizer V2, supporting partial and conditional replacement (#45574)

nickgg committed 3 years ago

19da1d22

[FakeLowP] T76913842 Make AddFakeFp16 take int inputs (#45992)

venkatacrc committed 3 years ago

a36f11a3

[JIT] Fix Dict bug in constant hashing (#45929)

Elias Ellison committed 3 years ago

c86655a8

[JIT] fix dict update (#45857)

Elias Ellison committed 3 years ago

72e4f51b

Revert D24093032: Improve logging in ProcessGroupNCCL for debugging purposes.

Natalia Gimelshein committed 3 years ago

de0d0bd5

[dist_optim] serialize compilation when creating dist_optim (#45871)

wanchaol committed 3 years ago

505be08c

Define objects using classes instead of namedtuples in torch.utils.data._utils.worker (#45870)

Loquats committed 3 years ago

ce82b522

[caffe2] Do not run RemoveOpsByType on recurrent networks (#45986)

Hao Lu committed 3 years ago

0927e02a

Improve logging in ProcessGroupNCCL for debugging purposes. (#45780)

pritamdamania committed 3 years ago

c8d76ff7

Parametrize # of longest tests in print_test_stats (#45941)

samestep committed 3 years ago

8fb32b9f

annotate torch.autograd.* modules (#45004)

guilhermeleobas committed 3 years ago

9679e1af

[quant] Add quantized Sigmoid module (#45883)

jerryzh168 committed 3 years ago

83d2c9a2

`torch.matrix_exp` doc fix (#45909)

nikitaved committed 3 years ago

30bf799f

Automatic update of fbcode/foxi to 6a4e19a2aaf7ae4b9fa9597526e65b395d5e79ad (#45951)

Yinghai Lu committed 3 years ago

b186831c

add test sharding to CUDA on linux (#45972)

janeyx99 committed 3 years ago

5a277370

[ONNX] Improve error handling for adaptive_pool (#45874)

neginraoof committed 3 years ago

5ce31b6f

Revert D20850851: caffe2/plan_executor: wait for 1 minute after exception and then abort

Rong Rong committed 3 years ago

1bb2d41b

Allow consumer ops to sync on GraphRoot's gradient (#45787)

mcarilli committed 3 years ago

5640b79b

Compress NVCC flags for Windows (#45842)

peterjc123 committed 3 years ago

bb99bea7

[JIT] Make objects throw Python AttributeError on nonexistant attr access (#45911)

James Reed committed 3 years ago

be45c340

[FX] Track use nodes in Node (#45775)

James Reed committed 3 years ago

8cdb638c

Newer Older