strisunshinewentingwang/pytorch

Pull Requests Commits

Fix include files for out-of-tree compilation (#48827)

CaoZhongZ committed 3 years ago

aff0b68a

Replace THError() check in THCTensorMathReduce.cu with C10_CUDA_KERNEL_LAUNCH_CHECK() (#49424)

Amogh Akshintala committed 3 years ago

16f4b0ed

[numpy] torch.{all/any} : output dtype is always bool (#47878)

kshitij12345 committed 3 years ago

c508e5b1

[JIT] Support multiple outputs in subgraph matcher. (#48992)

Mikhail Zolotukhin committed 3 years ago

38a59a67

[static runtime] refine fusion group (#49340)

bwasti committed 3 years ago

3ffe9e0f

[te] Fix bugs with shift operators (#49396)

bertmaher committed 3 years ago

f4e15c4a

Making ops c10-full: Generator arguments (#49013)

smessmer committed 3 years ago

5912316c

Making ops c10 full: out overloads with default arguments (#49012)

smessmer committed 3 years ago

a6274c12

Making ops c10-full: Dimname arguments (#49008)

smessmer committed 3 years ago

b47fa5e8

Making ops c10-full: ops blocked by manual registrations (#49007)

smessmer committed 3 years ago

c5f90a25

Making ops c10 full: ops returning multiple out arguments (#49006)

smessmer committed 3 years ago

e391dbc1

Make out ops c10-full (with hacky-wrapper) (#48912)

smessmer committed 3 years ago

40a02e2d

Suppress `warning: calling a constexpr __host__ function from a __host__ __device__ function is not allowed` warning (#49197)

r-barnes committed 3 years ago

11334280

[WIP][FX] Add FX page to docs (#48814)

James Reed committed 3 years ago

77800691

fix test_dispatch tests to error on duplicate def (#49254)

bdhirsh committed 3 years ago

9908b93d

Revert D25494735: Update TensorPipe submodule

Lucas Hosseini committed 3 years ago

8ae9b46e

Make WorkNCCL use CUDAEvent::query() rather than re-implement it (#49343)

lw committed 3 years ago

9234f502

Update TensorPipe submodule (#49232)

lw committed 3 years ago

5a5e576a

Do not return unitialized qschame from getQSchemeAndQParamVector (#49391)

malfet committed 3 years ago

98726119

Fix check_kernel_launches.py for macros and provide extended context (#49365)

r-barnes committed 3 years ago

39a10fb6

Revert D25135415: [PyTorch] Use plain old function pointer for RecordFunctionCallback

Mike Ruberry committed 3 years ago

25bc9062

Add assertion on any NaN error on the error feedback (#49374)

Yi Wang committed 3 years ago

a419a3e2

[PyTorch] Use plain old function pointer for RecordFunctionCallback (#48629)

swolchok committed 3 years ago

7e23ee15

[PyTorch] remove convenience RecordFunctionCallback interface (#48620)

swolchok committed 3 years ago

900aa4ee

Fix typo in torch.load docstring for the `f` parameter (#49350)

mzjp2 committed 3 years ago

bbeee481

[te] Ban uint8 tensors from fusion groups (#49247)

bertmaher committed 3 years ago

626b8c0c

Enable BF16 for indexing on CUDA (#48801)

zasdfgbnm committed 3 years ago

50b361a8

Fix Windows CUDA-11.1 test jobs (#49376)

malfet committed 3 years ago

23e98e73

Add Kernel Launch Checks to files under caffe2/aten/THC (#49358)

Amogh Akshintala committed 3 years ago

e2510a0b

[aten] index_select dim 1 (#47077)

ajyu committed 3 years ago

cb3169d7

Newer Older