strisunshinewentingwang/pytorch

Pull Requests Commits

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

[pytorch] Expand PixelShuffle to support any number of batch dims (#49187)

jbschlosser committed 3 years ago

220b9166

Use Unicode friendly API on Win32 in THAllocator (#47905)

skyline75489 committed 3 years ago

3a943e9f

Fixed cat transform to work with event_dim > 0 (#49111)

vikigenius committed 3 years ago

1e2d1d72

Add kernel launch checks in caffe2/aten/src/ATen/native/cuda/ (#49269)

Amogh Akshintala committed 3 years ago

d5a971e1

Add another way to verify ccache in CONTRIBUTING.md (#49337)

janeyx99 committed 3 years ago

86cf1e13

Revert D25489030: [PyTorch] Make tls_local_dispatch_key_set inlineable

ezyang committed 3 years ago

6820745e

Refactor: use version instead of major version in windows build (#49156)

mszhanyi committed 3 years ago

4188c374

Remove type annotations from signatures in html docs (#49294)

rgommers committed 3 years ago

6cfd7c38

sls + layernorm test (#43799)

Hector Yuen committed 3 years ago

9e3c25ff

[PyTorch] Make tls_local_dispatch_key_set inlineable (#49264)

swolchok committed 3 years ago

be849ed1

[CUDA graphs] Cuda RNG-safe graph capture and replay bindings (#48875)

mcarilli committed 3 years ago

c068180a

[CrashFix] Make the dst tensor contiguous when copying from metal

xta0 committed 3 years ago

25833e5d

[AARCH64] Fix vst1q_f32_x2 implementation (#49273)

Sebastian Pop committed 3 years ago

a0432a70

CUDA BF16 sparse (#48807)

zasdfgbnm committed 3 years ago

87636c07

Newer Older