strisunshinewentingwang/pytorch

Pull Requests Commits

fix INTERNAL ASSERT FAILED for maximum (#48446)

FNSTER committed 3 years ago

30324d1e

Fix AttributeError in _get_device_attr (#48406)

hauntsaninja committed 3 years ago

1c02be1b

fix move default not compile correctly on cuda92 (#48257)

Rong Rong committed 3 years ago

4fe583e2

add new build settings to torch.__config__ (#48380)

walterddr committed 3 years ago

54022e4f

Add cuda_ipc channel to TensorPipe (#46791)

mrshenli committed 3 years ago

d9c76360

Let JIT unpickler to accept CUDA DataPtr from read_record_ (#46827)

mrshenli committed 3 years ago

e3713ad7

centos now installs cmake from conda (#48035)

jeffdaily committed 3 years ago

5f181e2e

[PyTorch] Lazily construct guts of RecordFunction (#47550)

swolchok committed 3 years ago

3ceec73d

[PyTorch] Make RecordFunctionCallback::should_run_ a function pointer (#48274)

swolchok committed 3 years ago

d1df4038

change global_fp16_constants for test_fc_nnpi_fp16 (#48663)

Hector Yuen committed 3 years ago

9342b973

[ROCm] remove builds for versions less than 3.8 (#48118)

jeffdaily committed 3 years ago

c5f1117b

fix issue by which pytorch_jni is not bundled in libtorch (#46466)

Elijah Rippeth committed 3 years ago

aaf6582d

Remove `balance` and `devices` parameter from Pipe. (#48432)

pritamdamania committed 3 years ago

7c73fda5

Fuzzing benchmark for FFT operators (#47872)

peterbell10 committed 3 years ago

74d6a610

Fix complex tensors and missing data in benchmark utility (#47871)

peterbell10 committed 3 years ago

df6fc3d8

fx quantization: add option to leave graph inputs and/or outputs quantized (#48624)

vkuzo committed 3 years ago

f80aaadb

Revert D25172740: [pytorch][PR] [CUDA graphs] Make CUDAGeneratorImpl capturable

Mike Ruberry committed 3 years ago

98fddc1f

Add CUDA kernel checks to fbcode/caffe2/caffe2/sgd (#48347)

Kurtis David committed 3 years ago

0066b941

Change the type hints of "pooling.py". (#48412)

Xianguang-Zhou committed 3 years ago

736e8965

Revert D25222215: [quant][fix] Add bias once in conv_fused

malfet committed 3 years ago

c81f2d9a

Fix incorrect warnings in ParameterList/Dict (#48315)

CedricPicron committed 3 years ago

dc7ab46d

Add LazyConvXd and LazyConvTransposeXd (#47350)

asi1024 committed 3 years ago

492683bd

[vulkan] convolution old prepacking via cpu-shader (#48330)

IvanKobzarev committed 3 years ago

ccd20e99

[ONNX] Cast Gather index to Long if needed (#47653)

jiafatom committed 3 years ago

55fc0e9e

[ONNX] Support nonzero(*, as_tuple=True) export (#47421)

BowenBao committed 3 years ago

02e58aab

[caffe2][torch] Clean up unused variable 'device' (#48600)

jdonald committed 3 years ago

acd4fca3

Testing: Improve interaction between dtypes and ops decorators (#48426)

peterbell10 committed 3 years ago

9500e8a0

[quant][fix] Add bias once in conv_fused (#48593)

jerryzh168 committed 3 years ago

d2e42986

add aot_based_partition (#48336)

scottxu0730 committed 3 years ago

7a59a1b5

[Gradient Compression] Add a random generator to PowerSGD state for initializing low-rank matrix Q (#48507)

Yi Wang committed 3 years ago

ddb65949

Newer Older