strisunshinewentingwang/pytorch

Pull Requests Commits

[fix] torch.cat : cross-device check for out and input tensors (#53004)

kshitij12345 committed 3 years ago

fe38027f

compareSet method for HashStore and FileStore (#53803)

siva1007 committed 3 years ago

fdbd667e

Fix typo "informations" -> "information" (#53746)

lithuak committed 3 years ago

d4c877b5

Add missing decorators in test_spectral_ops (#53736)

peterbell10 committed 3 years ago

f62e9156

fix for method_tests() random failures (#53854)

aocsa committed 3 years ago

89fce74d

[caffe2] Support deserializing tensors using alternate serialization formats (#53403)

simpkins committed 3 years ago

33aaea91

[caffe2] add a CAFFE2_NODISCARD macro to help support old compilers (#53754)

simpkins committed 3 years ago

91531d30

Use the conda channel defined in docker.Makefile to install cudatoolkit (#53316)

xuzhao9 committed 3 years ago

7763bb6c

[CUDA graphs] Private mempools for CUDA graphs (#51436)

mcarilli committed 3 years ago

90dfdef2

[PyTorch] Remove unnecessary assert in maybe_resize_storage_cpu (#53724)

swolchok committed 3 years ago

804f3f98

Replace thrust with cub in randperm (#53841)

zasdfgbnm committed 3 years ago

34eb644e

Skip dispatch for is_signed (#53847)

Caroline Chen committed 3 years ago

7f4aff82

ns for fx: move linear activation test case to new API (#53777)

vkuzo committed 3 years ago

2912ad13

ns for fx: move compare activations for conv test to new API (#53776)

vkuzo committed 3 years ago

57bf1340

ns for fx: move lstm dynamic weight test case to new API (#53772)

vkuzo committed 3 years ago

01c6e936

ns for fx: move linear dynamic weight test case to new API (#53765)

vkuzo committed 3 years ago

a71cd135

ns for fx: move linear weight test case to new API (#53764)

vkuzo committed 3 years ago

9c8f112a

ns for fx: move conv weight test case to new API (#53761)

vkuzo committed 3 years ago

19fe8a52

ns for fx: extract common code in tests to util functions (#53748)

vkuzo committed 3 years ago

986e3c0a

ns for fx: clean up API naming (#53729)

vkuzo committed 3 years ago

7d27eb80

ns for fx: add support for logging inputs

vkuzo committed 3 years ago

421e91df

ns for fx: change dtype cast from once per N node to once per node

vkuzo committed 3 years ago

cc940f35

ns for fx: change API to take nn.Module instead of GraphModule (#53075)

vkuzo committed 3 years ago

d73e36a4

Fix run_test_module logic (#53884)

malfet committed 3 years ago

b00cdfe1

Do not use shards for single run tests (#53883)

malfet committed 3 years ago

ae7984b1

fix static dispatch linker error (#53859)

bdhirsh committed 3 years ago

a7ddd15d

[doc] reorg dist init and non-init functions (#52976)

stas00 committed 3 years ago

924c15c9

[DataLoader] ZipIterDataPipe (#53554)

ejguan committed 3 years ago

fff0a3f9

Add support for single tensor in `inputs` argument for backward (#53827)

soulitzer committed 3 years ago

7297556d

fx quant: clean up names of quantize handlers (#53614)

vkuzo committed 3 years ago

4884a6ab

Newer Older