strisunshinewentingwang/pytorch

Pull Requests Commits

Set USE_KINETO=1

ilia-cher committed 3 years ago

ed1508d6

Remove set_quantizer_ from native_functions.yaml (#49463)

smessmer committed 3 years ago

26974e6b

Revert D25574962: [pytorch][PR] Updated derivative rules for complex svd and pinverse

Mike Ruberry committed 3 years ago

f5b68e74

[pt] fuse ClipRangesGatherSigridHash (#49181)

ajyu committed 3 years ago

c18af03a

Adding fix for invalid annotation types for dictionary (#49425)

nikithamalgifb committed 3 years ago

26e076d1

Change aten::native_layer_norm signature to match torch.layer_norm definition (#48971)

rdspring1 committed 3 years ago

65876d3f

Add BFloat16 support for isinf and isfinite (#49356)

zasdfgbnm committed 3 years ago

2ea1d97e

[quant][be] Add typing for quantization_mappings.py (#49179)

jerryzh168 committed 3 years ago

ede0b169

Bring back math_silu_backward which works for all backends. (#49439)

Ailing Zhang committed 3 years ago

4edaf4d7

Add torch._foreach_zero_ API (#47286)

Iurii Zdebskyi committed 3 years ago

6230e337

Set caffe2::pthreadpool() size in ParallelOpenMP (#45566)

dbalchev committed 3 years ago

4ce2b0b0

[reland] Support torch.distributed.irecv(src=None, ...) (#49383)

pritamdamania committed 3 years ago

db2ecefc

add files to SLOW_TESTS for target determinator (#49500)

Rong Rong (AI Infra) committed 3 years ago

df233709

fx quant: make sure observer is inserted before a quantized output (#49420)

vkuzo committed 3 years ago

82ac6c75

fx quant: fix fq when input is quantized and node does not need fq (#49382)

vkuzo committed 3 years ago

84506e03

fx quant: do not insert observers at quantized inputs (#49239)

vkuzo committed 3 years ago

75420760

fx quant: move {input|output}_quantized_idxs cfg from convert to prepare (#49238)

vkuzo committed 3 years ago

92df8706

eager quant: remove fake_quant after add/mul nodes during QAT (#49213)

vkuzo committed 3 years ago

36b20923

Add fusion support of aten::to (#48976)

Elias Ellison committed 3 years ago

90458627

[NNC] add support for masked_fill (#48974)

Elias Ellison committed 3 years ago

80b508f2

[NNC] Add Support For is_nan (#48973)

Elias Ellison committed 3 years ago

50386b99

[extensions] fix `is_ninja_available` during cuda extension building (#49443)

stas00 committed 3 years ago

60b4c401

Fix CUDA extension ninja build (#49344)

zasdfgbnm committed 3 years ago

d409da06

Relax the atol/rtol of layernorm math kernel test. (#49507)

Ailing Zhang committed 3 years ago

1c6e179b

Fix bad error message when int overflow (#48250)

Kiyosora committed 3 years ago

c675727a

.circleci: Only downgrade if we have conda (#49519)

seemethere committed 3 years ago

a5cc0a6f

Prevent accidentally writing old style ops (#49510)

smessmer committed 3 years ago

872f6486

[NNC] Dont inline outputs buffers on cpu (#49488)

Elias Ellison committed 3 years ago

9056173a

Revert D25569586: stft: Change require_complex warning to an error

Mike Ruberry committed 3 years ago

47c65f82

Introduce tools.codegen.api.translate (#49122)

ezyang committed 3 years ago

3efd5d8f

Older