strisunshinewentingwang/pytorch

Pull Requests Commits

[JIT] Add GRAPH_DEBUG for conv_add_relu fusion (#73138)

davidberard98 committed 2 years ago

8af23ea0

Revert D33994011: Make debug_pkl smaller by only emitting unique traces.

albanD committed 2 years ago

3bd1507f

[Static Runtime] Avoid checks during op execution for TupleConstruct & ListConstruct (#69029)

d1jang committed 2 years ago

5772b1af

Revert D34400588: [pytorch][PR] super setUp call missing in TestSparse

albanD committed 2 years ago

49444bb5

Introduce an environment variable to change c10 log level (#71746)

cbalioglu committed 2 years ago

7366724e

Make debug_pkl smaller by only emitting unique traces. (#72596)

qihqi committed 2 years ago

3d37f5b0

Check clang++/g++ version when compiling CUDA extensions (#63230)

andfoy committed 2 years ago

86deecd7

Update ONNX submodule to 1.11.0 (#73111)

garymm committed 2 years ago

3ac78281

Move Lazy Shape Inference functions to pytorch core (#72756)

wconstab committed 2 years ago

d56d530e

[torch] do not fold bmm into mm when tensor1 dim==3 but not contiguous (#73115)

jspark1105 committed 2 years ago

fffb97f3

[PyTorch] [Model Tracer] Protect writes to set/map for capturing used operators, dtypes, custom classes, and build features (#73327)

dhruvbird committed 2 years ago

f4611ee9

[PyTorch] [Model Tracer] Refactor method bodies in various record function handlers into source (.cpp files) (#73321)

dhruvbird committed 2 years ago

c2279d48

Enable CSR inputs for torch.sparse.mm (#73075)

IvanYashchuk committed 2 years ago

ebd93f69

Add dynamic shape support to AOT driver & compiler (#72995)

priyaramani committed 2 years ago

ac97e953

Squeeze p2: hook up Squeeze to LazyView (#73067)

Krovatkin committed 2 years ago

5a7778c9

[shard] fix some imports in tests (#73309)

wanchaol committed 2 years ago

6c6ae0e7

[shard] Extensible ShardingSpec (#72130)

wanchaol committed 2 years ago

d6c5295e

[vulkan] Remove dead code from previous Vulkan backend (#73243)

SS-JIA committed 2 years ago

b3e3eb99

Add BUILD_LAZY_CUDA_LINALG option (#72306)

malfet committed 2 years ago

82130758

Make "server socket not listening" warning logs less noisy (#73149)

cbalioglu committed 2 years ago

fc3c7fb7

Introduce debug and trace log levels in c10d (#73167)

cbalioglu committed 2 years ago

e143f980

Refactor TORCH_DISTRIBUTED_DEBUG implementation (#73166)

cbalioglu committed 2 years ago

e1db2f13

Prefix c10d log messages with `[c10d]` for easier troubleshooting (#73144)

cbalioglu committed 2 years ago

3d9706c4

[GH1] Support merging PRs with more than 100 files

malfet committed 2 years ago

ec493730

[GH1][BE] Move GQL mocked data to json db

malfet committed 2 years ago

e3019a96

[Static Runtime] Enable all tests to run with TensorExpr fuser (#73263)

navahgar committed 2 years ago

4838c6dc

Fix `nn.Module.state_dict()` (#72780)

lkct committed 2 years ago

9d6639ab

Move vectorized CPU codegen to after ATen codegen (#72869)

ezyang committed 2 years ago

2321f26f

[pytorch][nn] torch.nn.MultiheadAttention fix (#73053)

Huaqing Xiong committed 2 years ago

eabb0a08

[tensorexpre] Add typecast when src and dest buf types are different in PlacementAllocate (#71934)

huiguoo committed 2 years ago

5eb5b612

Newer Older