strisunshinewentingwang/pytorch

Pull Requests Commits

nvprims native batch norm patch (#88455)

jjsjann123 committed 1 year ago

52375a0f

[Dynamo] Improve BuiltinVariable log when incorrect arg count happens (#88409)

yanboliang committed 1 year ago

b1116a51

Fix minifier accuracy msg (#88515)

mlazos committed 1 year ago

5220d07d

Populate self.export in InstructionTranslatorBase (#88508)

mergennachin committed 1 year ago

dde9affe

[QNNPACK] Add unaligned attributes where asan fails (#88276)

digantdesai committed 1 year ago

afdc2283

[Quant] Respect non_leaf_module_list for activation modules (#88498)

andrewor14 committed 1 year ago

7560a7b2

[BE] Update native_functions.yaml README; we do not support Tensor! (#88513)

janeyx99 committed 1 year ago

5af3feef

Support DDP ignored parameters in DDPOptimizer (#88460)

wconstab committed 1 year ago

678d0380

enable backward for log1p (sparse layouts) (#88155)

amjames committed 1 year ago

ff6770a9

Support sparse inputs to deg2rad (#88156)

amjames committed 1 year ago

6938dd0b

Enable sparse_csr autograd testing for relu (#88154)

amjames committed 1 year ago

1964d8c3

Add sparse layout support for torch.frac (#88153)

amjames committed 1 year ago

f03302ba

Disable mem leak check (#88373)

clee2000 committed 1 year ago

d632d94c

Re-enable inductor models tests as periodical jobs (#88509)

huydhn committed 1 year ago

093e2208

Don't print fatal:... in generate_torch_version.py (#88335)

janeyx99 committed 1 year ago

3e6579b8

[inductor] Handle the case where kwargs contains tensor (#88417)

desertfire committed 1 year ago

955cbe61

Add nondeterministic error for `scatter` (#88244)

kurtamohler committed 1 year ago

e940a2f8

[fx2ait] fixes for AITSplitter (#87805)

mortzur committed 1 year ago

6575174d

[NVFuser] Upstream push 1026 (#87779)

jjsjann123 committed 1 year ago

7b419e85

[MPS] Fix embedding backward with scalar index (#82809)

qqaatw committed 1 year ago

15e54293

Modified roundup_power2_divisions to specify the number of divisions for each power of two interval (#87290)

c-odrin committed 1 year ago

5b767d40

[vulkan] enable prepacking for Batchnorm op (#88433)

SS-JIA committed 1 year ago

b78b8727

Revert "Revert "Put Python Dispatcher cache in dict, clear it on new registrations. (#88329)"" (#88489)

ezyang committed 1 year ago

53eac1d4

nvprim python runtime dtype correctness patch (#88452)

jjsjann123 committed 1 year ago

79abea56

Revert "remove assert_allclose from torch.testing (#87974)"

pytorchmergebot committed 1 year ago

8c1c6759

Fix typo in clones (#88501)

ezyang committed 1 year ago

bda688c1

[torch package] Treat builtins as default extern module (#88385)

842974287 committed 1 year ago

633f0d62

Add dep on Accelerate framework to torch podspecs (#88422)

jmdetloff committed 1 year ago

ead36e5a

[Vulkan][TCC] Add tests for conv2d prepack context (#88316)

manuelcandales committed 1 year ago

dc00bb51

Add use_lazy_shape flag to GenLazyIr class (#88444)

wonjoolee95 committed 1 year ago

a171b063

Newer Older