strisunshinewentingwang/pytorch

Pull Requests Commits

Implement Observer pass on simple model and validate stats (#18848)

nishantpdce committed 5 years ago

08ee4e56

AVX2 with GCC9 fix. (#18991)

cbalint13 committed 5 years ago

67fdb4ab

Remove tensorFromBlob() from Type (#18779)

Roy Li committed 5 years ago

f6af76ea

Improve precision of emitted code for prim::Constant (#18817)

James Reed committed 5 years ago

9b69f21a

convert_sync_batch_norm to SyncBatchNorm (#18787)

iArunava committed 5 years ago

79533ef0

fix bug when falling back to acc32 when weight is prepacked (#18974)

csummersea committed 5 years ago

907b4c58

move 2ops back to autodiff (#18969)

Ailing Zhang committed 5 years ago

dbd9971d

Preserve naming for inputs/outputs with observer insertion (#18713)

nishantpdce committed 5 years ago

1ffa358f

Emit math functions specific to output type (#18815)

James Reed committed 5 years ago

34382e42

add instructions for NVIDIA Jetson platforms (#18990)

soumith committed 5 years ago

8961ad8c

Quantizer pass to insert quant-dequant nodes into IR (#18446)

nishantpdce committed 5 years ago

bcd52719

add SyncBatchNorm to docs (#18988)

soumith committed 5 years ago

7b5b1486

Add c10_cuda to libraries in CUDAExtension for Windows (#18982)

mooncake4132 committed 5 years ago

d6d0fcc9

Remove Trainer from README.md (#18980)

zasdfgbnm committed 5 years ago

1497d453

Create Object that represents a Module (#18469)

zdevito committed 5 years ago

13f03a42

Add numpy like repeat as torch.repeat_interleave (#18395)

zasdfgbnm committed 5 years ago

8c9caf18

Fix interpolate trace (#18875)

Elias Ellison committed 5 years ago

e6bbbb01

Code string API for fuser testing (#18884)

James Reed committed 5 years ago

60849082

remove unused func (#18712)

suo committed 5 years ago

ce67775f

Revert D14778810: [caffe2/int8] fix bug when falling back to acc32 when weight is prepacked

bddppq committed 5 years ago

46fe2665

slots with explicit value/setValue make more sense in future patches (#18468)

zdevito committed 5 years ago

f6f34b3f

Make Object hold its ClassType (#18467)

zdevito committed 5 years ago

091acb09

Enforce single parent for script submodules (#18379) (#18860)

zdevito committed 5 years ago

53458c97

CUDA_NVCC_EXECUTABLE is not needed, as nvcc is in PATH (#18958)

stas00 committed 5 years ago

f9a56d4a

Fix precision issue with expansion that prefers 'probs' over 'logits' (#18614)

ahmadsalim committed 5 years ago

8e1e2912

Method is supposed to be in-place (#18684)

NegatioN committed 5 years ago

b90cbb84

fix bug when falling back to acc32 when weight is prepacked (#18881)

csummersea committed 5 years ago

28990f34

More numerically stable lerp (#18871)

mkolod committed 5 years ago

c1790fa2

Increase default c10d/ProcessGroupGloo test timeout (#18916)

pietern committed 5 years ago

edc7b472

remove symbolic variable part 1 (#17986)

Ailing Zhang committed 5 years ago

cb3a4a3d

Newer Older