strisunshinewentingwang/pytorch

Pull Requests Commits

davidberard98 committed 1 year ago

e4e30829

Set interface_call to true be default (#86668)

qihqi committed 1 year ago

f32aeeae

[Experimentation] Add TSAN build and test (#85313)

huydhn committed 1 year ago

7f02f2ac

Optimize __dlpack_device__ performance (#86665)

huww98 committed 1 year ago

92562046

[nn] Add remove_duplicate flag to named_buffers (#674) (#85903)

jerryzh168 committed 1 year ago

c12f829c

Docs: fx.Node docs incorrectly state that the self argument is included in args for module calls (#86685)

cherrywoods committed 1 year ago

693250ac

Add test case for matrix multiply-add with large inputs (#85550)

fangwangcn committed 1 year ago

160118d7

Fix torch histogramdd docstring (#86593)

vfdev-5 committed 1 year ago

212fa874

[BE] Fix python docs typos up till torch.chunk (#86642)

janeyx99 committed 1 year ago

f26292d9

min/max support for SymInt/Floats, finish as_strided/scatter/squeeze() backward symint support (#86643)

albanD committed 1 year ago

86f914e9

Add module: decompositions as an owner to test_decomp.py (#86703)

janeyx99 committed 1 year ago

6923dc3b

Move functorch tests from functorch/test/* to test/functorch/* (#86623)

zou3519 committed 1 year ago

109f4d44

Upcast to fp32 in test_addmm_block ref_half_bfloat16 (#86682)

IvanYashchuk committed 1 year ago

51ea4418

Revert "Add meta support for _adaptive_avg_pool2d_backward (#86359)"

pytorchmergebot committed 1 year ago

3edf79dc

Improve interpolate() speed for channels_last CPU images and masks (#86361)

NicolasHug committed 1 year ago

97de2811

Pin numpy version during MPS tests (#86691)

malfet committed 1 year ago

a4ee6956

[CUBLAS][CUDA GRAPHS] (re-re-re-re-open of #83461) Explicitly set the workspace for cuBLAS handles (#86645)

eqy committed 1 year ago

352d9264

Add version selector back to functorch docs (#86602)

zou3519 committed 1 year ago

937d677d

Add meta support for _adaptive_avg_pool2d_backward (#86359)

anjali411 committed 1 year ago

a56a8c0f

Skip forward AD tests for torch.native_batch_norm (#86206)

IvanYashchuk committed 1 year ago

03d8ab4d

[FSDP][Easy] Rename `_prefixed_param_names` -> `_fqns` for consistency (#86653)

awgu committed 1 year ago

6ab07feb

Symintify NLL loss, copy and squeeze (#86606)

albanD committed 1 year ago

2fe58085

More symintification of get/set item (#86605)

albanD committed 1 year ago

be862782

symintify autograd view chaining (#86604)

albanD committed 1 year ago

f8414422

symintify einsum (#86603)

albanD committed 1 year ago

49c9b0a1

Revert "Improve interpolate() speed for channels_last images and masks (#86361)"

pytorchmergebot committed 1 year ago

3a2cfbb8

index op with int32 support (#86318)

jianyuh committed 1 year ago

17074389

fix: half reduction with multiple sub-iterators (#85596)

kshitij12345 committed 1 year ago

88a8a900

Enable capturing of comm collective parameters (#98) (#85368)

louisfeng committed 1 year ago

55479fe8

[torchdynamo hash update] update the pinned torchdynamo hash (#86651)

pytorchmergebot committed 1 year ago

ad2b04c3

Older