strisunshinewentingwang/pytorch

Pull Requests Commits

Silu support Complex for CUDA (#106854)

FFFrog committed 1 year ago

871d7d24

fixing internal test failure on non sm_80 machines (#107340)

HDCharles committed 1 year ago

3ddf3050

[vision hash update] update the pinned vision hash (#107498)

pytorchupdatebot committed 1 year ago

b5642f0b

[pytorch][fakepg] enhance fakepg: broadcast and scatter (#107480)

xmfan committed 1 year ago

302278b4

Update reduction_ops groupings to include primtorch types (#107338)

ekamiti committed 1 year ago

017499b0

Update submodule NCCL to v2.18.3 (#104993)

Skylion007 committed 1 year ago

93f2a64d

Revert "Use CUDA DSA in ATen (#95300)" (#107483)

ezyang committed 1 year ago

64e02de9

Update shape_funcs to test primtorch operators (#107336)

ekamiti committed 1 year ago

2d7a062d

Revert "Revert "Reland "Add forward mode AD to out-place foreach functions (#102409) (#106043)""" (#106320)

crcrpar committed 1 year ago

5814380e

[ROCm] Update ROCm skip decorators (#106138)

lcskrishna committed 1 year ago

bc662fff

[quant][pt2e] Move specific quantizer related things outside of main quant code base (#106806) (#107259)

jerryzh168 committed 1 year ago

28be2c67

Remove jbschlosser from symbolic-shapes auto request list (#107482)

ezyang committed 1 year ago

4ee62247

Enable mypy check in torch/_inductor/fx_passes/post_grad.py (#107449)

FFFrog committed 1 year ago

35e222e1

[pt2] test if core decomps are differentiable (#107241)

nkaretnikov committed 1 year ago

77f080ee

Update binary_ufuncs groupings to include primtorch types (#107419)

ekamiti committed 1 year ago

5b7b9e78

Change >= in the GRU and the LSTM document to \ge (#107379)

bitsandscraps committed 1 year ago

af0ed25e

Enable mypy check in torch/_inductor/kernel/unpack_mixed_mm.py (#107445)

FFFrog committed 1 year ago

c2706e5b

add more check on LSTMCell (#107380)

FFFrog committed 1 year ago

2d2d43d9

[Compiled Autograd] Fix duplicate visits of same node (#105887)

jansel committed 1 year ago

bdecdfd2

Add verbose_guards logging artifact (#107388)

ezyang committed 1 year ago

67bb3c05

Add fast traceback utilities (#107358)

ezyang committed 1 year ago

36bb7a1f

Hand bind CapturedTraceback (#107438)

ezyang committed 1 year ago

d5f7df3b

[dtensor][1/n] refactor op dispatch logic to reduce overhead (#107305)

wanchaol committed 1 year ago

d8f2ef10

[dynamo] Make KeyedJaggedTensor a variable. (#107319)

zhxchen17 committed 1 year ago

8d6a487d

Make StarDep.index throw an error (#107092)

jansel committed 1 year ago

ea3381d9

Make Openxla dynamo backend take boxed input (#107260)

JackCaoG committed 1 year ago

139437bb

Revert "Fakify leaf of FunctionalTensor (#107062)"

pytorchmergebot committed 1 year ago

3c11184c

Inductor cpp wrapper: fix optional tensor input (#106847)

chunyuan-w committed 1 year ago

c21e9de2

enable mkl_gemm_bf16bf16f32 in cpublas::gemm (#107196)

mingfeima committed 1 year ago

e10791c0

reseed all Generators in Dataloader's _worker_loop() -- via GC (#107131)

NicolasHug committed 1 year ago

42625da5

Newer Older