strisunshinewentingwang/pytorch

Pull Requests Commits

[Quant] Improve performance of ONEDNN backend (#84470)

Xia-Weiwen committed 2 years ago

6c48a01c

Revert "Python stack tracing OD flow (part 1) (#84362)"

pytorchmergebot committed 2 years ago

35088f28

[vision hash update] update the pinned vision hash (#85315)

pytorchmergebot committed 2 years ago

8c7e2097

[torch.futures] Fix nullptr deref (#85304)

malfet committed 2 years ago

c05ca0db

[functorch] Fix dangling impls (#85299)

zou3519 committed 2 years ago

66907e72

Revert "Reduce memory usage requirement of `test_warp_softmax_64bit_indexing` in `test_nn.py` (#85037)"

pytorchmergebot committed 2 years ago

53fdd606

Fix segfault for `out` with a large number of dims (#85294)

ngimel committed 2 years ago

a998a8eb

Setup torch/csrc/functorch/*; move CompileCache.{h, cpp} there (#85263)

zou3519 committed 2 years ago

d9024ea2

Python stack tracing OD flow (part 1) (#84362)

slgong-fb committed 2 years ago

1f4f05e5

Reduce memory usage requirement of `test_warp_softmax_64bit_indexing` in `test_nn.py` (#85037)

eqy committed 2 years ago

66a9cba2

Handle implicit real->complex casting for backward of stack (#84993)

t-vi committed 2 years ago

e41d758e

Add aten _assert_tensor_metadata op (#84617)

voznesenskym committed 2 years ago

cd7408e9

Revert "Legalize BFloat16 in NNC. (#83988)"

pytorchmergebot committed 2 years ago

6ed90379

Fix exception handling, improve overheads and avoid constructing storage for element size (#84612)

Aidyn-A committed 2 years ago

1456cca1

[PolishComment] Polish code comment, revelant->relevant (#85238)

jiahy0825 committed 2 years ago

cbe5469e

Fix segfault case for torch.ormqr (#85278)

IvanYashchuk committed 2 years ago

8c952db1

Check that groups is > 0 in _convolution op (#85111) (#85248)

Thytu committed 2 years ago

555bb6cd

Revert "Land "Make ceil,floor,round,trunc handle integers" (#85144)"

pytorchmergebot committed 2 years ago

7234eb06

Fix bugs in sparse compressed tensor shape and device inference (#85240)

pearu committed 2 years ago

f0b06c64

Support for sym_strides() in backwards formulas (#85210)

ezyang committed 2 years ago

6a186162

When tracing SymInts, peephole optimize multiply by one (#85261)

ezyang committed 2 years ago

f38f9dfb

[NNC] support aten::_convolution when it is 2D conv (#84038)

chunyuan-w committed 2 years ago

ebf45a07

Legalize BFloat16 in NNC. (#83988)

EikanWang committed 2 years ago

b049493e

Land "Make ceil,floor,round,trunc handle integers" (#85144)

lezcano committed 2 years ago

b27eb8d3

Stop monkeypatching Tensor.backward() on `import functorch` (#85152)

zou3519 committed 2 years ago

cd32a86b

Stop loading jit decomps in eager_transforms.py (#85151)

zou3519 committed 2 years ago

5ce56d93

Make addmm meta kernel consistent with mm (#84960)

antoniojkim committed 2 years ago

6fd8e28a

Added docs and opinfo for narrow_copy (#84493)

srossross committed 2 years ago

3a51b557

[functorch] add batch rule for linalg.lu_solve (#85175)

zou3519 committed 2 years ago

b0c447e9

Adds normal prim, randn reference, and randn OpInfo (#85128)

Mike Ruberry committed 2 years ago

d561aa94

Newer Older