strisunshinewentingwang/pytorch

Pull Requests Commits

Add a ZT fastpath for linalg_cross (#76940)

Nikita Karetnikov committed 2 years ago

fc4f8b5e

Enable fp32/bf16 PRelu forward and backward in MkldnnCPU path (#60427)

yanbing-j committed 2 years ago

cd33e412

add BFloat16 support for UpSample on CPU

mingfeima committed 2 years ago

8d4e069e

[fix] reflection_pad3d: correct dispatch name (#77153)

kshitij12345 committed 2 years ago

e383b135

Fix math notation for linear projections in nn.RNN docs (#77082)

yashika-git committed 2 years ago

182b28fe

Faster `index_select` for sparse COO tensors on CPU. (#72710)

nikitaved committed 2 years ago

00a1fb64

[FX qconfig] add weighted_op_qint8_dtype_config for int8 TRT and import linear config to get_tensorrt_backend_config_dict (#76877)

brad-mengchi committed 2 years ago

d7035c1c

speed up 1d sort (#77100)

ngimel committed 2 years ago

ae25d14f

[DataPipe] Revamp serialization logic of DataPipes

NivekT committed 2 years ago

a008d19f

[CUBLAS][TF32] Add environment variable to allow override of `allow_tf32_cublas` (#77114)

eqy committed 2 years ago

cd916fea

Revert "Faster `index_select` for sparse COO tensors on CPU. (#72710)"

pytorchmergebot committed 2 years ago

8d67972b

Remove misleading statement in optim.Optimizer docs (#76967)

daniellepintz committed 2 years ago

6ae047b0

[ROCm] rccl performance improvement via env var (#76985)

jeffdaily committed 2 years ago

614e0459

Remove deprecated torch.solve (#70986)

IvanYashchuk committed 2 years ago

890bdf13

Revert "Sparse CSR: enable autograd for torch.sparse.addmm and torch.sparse.mm"

pytorchmergebot committed 2 years ago

f94abd59

[FSDP] Do not clone buffers; offload buffers to CPU if needed

awgu committed 2 years ago

9903f1ae

[FSDP] Change default auto wrap policy name (#76858)

zhaojuanmao committed 2 years ago

3621462e

[FSDP] Move param/buffer name comp. to ctor for `ignored_modules`

awgu committed 2 years ago

75f316f1

Disable ios 12.5.1 job as its logs don't reveal why it's failing (#77162)

janeyx99 committed 2 years ago

da3b8309

Minor torchhub docs

NicolasHug committed 2 years ago

da3ebaeb

Sparse CSR: enable autograd for torch.sparse.addmm and torch.sparse.mm

IvanYashchuk committed 2 years ago

721a8ca6

[JIT] fix opinfo utils to handle tensor kwargs

davidberard98 committed 2 years ago

41767672

[NVFuser] Opinfos for extremal values in binary ufuncs

davidberard98 committed 2 years ago

949cbf1d

RPC Meta device support

pbelevich committed 2 years ago

2e2200d7

ROCM: Increase timeout for flaky test_with_kwargs (#76706)

arindamroy-eng committed 2 years ago

24372eb5

Latest linalg_lu revert caused bc to fail, add exceptions (#77112)

janeyx99 committed 2 years ago

0272aa1f

Fix docker builds, cleanup cuda 115 (#77086)

atalman committed 2 years ago

8a6856ae

make has_bundled_input work for flatbuffer (#76854)

qihqi committed 2 years ago

41ff6f8c

Skip lu_solve meta tests (#77110)

malfet committed 2 years ago

b4fa1e88

disabling squeeze/unsqueeze; disabling BN/BN_BWD for perf concern (#77017)

jjsjann123 committed 2 years ago

489818e7

Newer Older