strisunshinewentingwang/pytorch

Pull Requests Commits

SubgraphMatcher: matching modules support. (#25075)

Mikhail Zolotukhin committed 5 years ago

85bca16a

SubgraphMatcher: add logging.

Mikhail Zolotukhin committed 5 years ago

16289c2f

SubgraphMatcher: Factor out matchAttributes.

Mikhail Zolotukhin committed 5 years ago

b5096b68

use avx2 for Add without broadcast and when inputs are uint8_t (#25098)

jspark1105 committed 5 years ago

a54f8f0f

Use the EmbeddingLookup API which takes the offsets instead of lengths (#24945)

jianyuh committed 5 years ago

363655dc

print padding_mode for Conv modules if not zeros (#23996)

Ailing Zhang committed 5 years ago

35a00155

Support lowering of fp16 weights

Yinghai Lu committed 5 years ago

add57fd2

Revert "per channel quantization support (#24936)" (#25131)

zrphercule committed 5 years ago

bc83ed10

move some methods into function.cpp (#25119)

suo committed 5 years ago

98544355

Add a skip_override option to should_run_job.py (#25118)

suo committed 5 years ago

65beee58

add some sparse tensor ops support in TorchScript (#24967)

wanchaol committed 5 years ago

5fd3251c

Add missing functions and methods for channelwise quantization (#24934)

dskhudia committed 5 years ago

12ea1d74

per channel quantization support (#24936)

dskhudia committed 5 years ago

9e996503

Per Channel quantization APIs (#24935)

dskhudia committed 5 years ago

aba15ce9

Make EmbeddingLookup APIs take offsets rather than lengths to match the PyTorch's EmbeddingBag (#24944)

jianyuh committed 5 years ago

ad7250d3

Update QNNPACK submodule to 901e9d4 (#25044)

dreiss committed 5 years ago

6981b4e5

Add helper function Tensor::names() (#24914)

zou3519 committed 5 years ago

c013c066

Rename Tensor::names() to Tensor::opt_names() (#24907)

zou3519 committed 5 years ago

530db2c7

Fix `FIXME_default_names` by storing static list of 64 none names (#24885)

zou3519 committed 5 years ago

867d8af2

Optimize performance for unboxed-only kernels (#25055)

smessmer committed 5 years ago

6c834246

Fix lint checker breakage caused by #25111 (#25122)

zrphercule committed 5 years ago

5b84514a

fix clang-tidy failing on master (#25121)

suo committed 5 years ago

199e15fa

dictPop: dereference dict.find() iterator before calling dict.erase() (#25056)

Basil Hosmer committed 5 years ago

2ec23804

fix annotated assignment (#25094)

Elias Ellison committed 5 years ago

ab38059b

Clean up after running doc tests (#25036)

davidriazati committed 5 years ago

1c4495d8

fix clang-tidy failing all the time on random lines (#25078)

suo committed 5 years ago

d1f0823d

Turn off fbgemm for libtorch android build (#25113)

IvanKobzarev committed 5 years ago

2cccad2c

pin_memory thread now uses 1 thread only (#25111)

ssnl committed 5 years ago

e42b238f

Add thread-local-state NamesMode and NoNamesGuard (#24942)

zou3519 committed 5 years ago

9a793a49

Fix python lints for generate_test_torchscripts.py (#25107)

IvanKobzarev committed 5 years ago

56245ffe

Newer Older