strisunshinewentingwang/pytorch

Pull Requests Commits

Optimize pytorch layer_norm forward (#20345)

xiaomengy committed 5 years ago

c9da0119

use tensoriterator instead of th for fill_ implementation. (#20719)

nairbv committed 5 years ago

9cec8ae1

Fix copy_transpose_valid check (#20759)

colesbury committed 5 years ago

7a0c6d52

make magic methods work with casts too (#20654)

Elias Ellison committed 5 years ago

5acc664f

Change Bias to QTensor with qint32(int32_t) (#20713)

jianyuh committed 5 years ago

e6f22e1b

Change Weight to QTensor with qint8(int8_t) (#20712)

jianyuh committed 5 years ago

b9a150ed

Fix a bug in quantize_linear (#20711)

jianyuh committed 5 years ago

ac2314fd

Update Conda description in PyTorch README (#20726)

dreiss committed 5 years ago

32803b52

Auto-convert GPU arrays that support the __cuda_array_interface__ protocol (#20584)

madsbk committed 5 years ago

5d8879cf

Improve the recommended citation

xuhdev committed 5 years ago

847d9c57

Add support for CMake switches for VS 2019 (#20752)

peterjc123 committed 5 years ago

bb20956e

add str comparisons (#20761)

Elias Ellison committed 5 years ago

47dc65fe

Add dequantize_linear for JIT pass (#20107)

jerryzh168 committed 5 years ago

cca923c4

Throw error if multiple kernels registered (#20737)

smessmer committed 5 years ago

cc02a1af

Hipify fb/quantize

xw285cornell committed 5 years ago

f3d827f3

Split cpu/gpu in caffe2/distributed + some clean up (#20674)

xw285cornell committed 5 years ago

b5edeca3

compile with -fcolor-diagnostics (#20662)

suo committed 5 years ago

d7cd2d7a

Fix missing cudatoolkit dependency in binary linux tests

pjh5 committed 5 years ago

c790f10e

Fixing upload_binary_htmls again

pjh5 committed 5 years ago

e3970d66

Revert D15178352: [pt1][quant] Quantized Conv2d operator

pjh5 committed 5 years ago

fac307a5

Fix -Wattributes warning on older versions of gcc (#20587)

nairbv committed 5 years ago

eca7fa35

Fixing missing miniconda path in macos smokes

pjh5 committed 5 years ago

712c60f9

Quantized Conv2d operator (#20064)

dskhudia committed 5 years ago

29b1b594

Add mandatory ScalarType nodes as input to the quant-dequant nodes. (#20468)

nishantpdce committed 5 years ago

d73caca2

Increase static tolerance for negative feature ids

Abhijith Reddy committed 5 years ago

371cf109

fix soft_nms_cpu call in BoxWithNMSLimit (#20738)

wat3rBro committed 5 years ago

0beecbda

Move bucketize_op to open source

olittle committed 5 years ago

fbdafdff

Refactor CUDA copy and general copy dispatch (#20685)

colesbury committed 5 years ago

320c3855

Add onnxifi support for Int8FCDNNLowPPackedWeightBlob (#20564)

Yinghai Lu committed 5 years ago

cf7ef5e6

restore hidden visibility by default for Linux builds (#20461)

kostmo committed 5 years ago

0bfc0eee

Newer Older