strisunshinewentingwang/pytorch

Pull Requests Commits

actually use reshape

samdow committed 2 years ago

92d66f25

fix error message

samdow committed 2 years ago

d9ddfb60

add decomposition for nll_loss2d_backward

samdow committed 2 years ago

82b6c9fd

[chalf] update type promotion table (#76893)

kshitij12345 committed 2 years ago

00fb8282

Move `contiguous_strides` to `c10/util/strides.h`.

ysiraichi committed 2 years ago

2e6ed5d0

replace "grep" with Python processing in `collect_env.py` (#77148)

pmeier committed 2 years ago

635aaa3d

torch/deploy,package: log usage for InterpreterManager, PackageExporter, PackageImporter (#77097)

d4l3k committed 2 years ago

4d30ebc8

[ROCm] update cmake package DIR paths (#77087)

jeffdaily committed 2 years ago

64b54343

shard `linux-xenial-py3.7-clang7-asan / test (default` from 4 to 5 (#77084)

clee2000 committed 2 years ago

ecab4fb0

Revert "speed up 1d sort (#77100)"

pytorchmergebot committed 2 years ago

54dbfc1e

faster batch sampler (#76951)

木澄 committed 2 years ago

d22d749a

Fix `broadcast_in_dim` support in NVFuser Frontend (#76790)

kevinstephano committed 2 years ago

752d496c

Parallelize `gatherTopK` on multiple blocks (#74267)

zasdfgbnm committed 2 years ago

8560fa73

Add meta tensor support for some operations using python registration

anjali411 committed 2 years ago

767af8e3

Profiling range for FSDP.forward (#76899)

rohan-varma committed 2 years ago

e31b6213

Fix max pool forward nhwc (#76597)

Aidyn-A committed 2 years ago

a127c584

Add __all__ for torch.autograd.{anomaly_mode, gradcheck, forward_ad}

anjali411 committed 2 years ago

65a8f8f6

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

Older