strisunshinewentingwang/pytorch

Pull Requests Commits

[quant][graphmode] use RemoveMutation to remove append (#41161)

jerryzh168 committed 4 years ago

4b4184fc

Op to create quant scheme blob (#40760)

Anurag Gupta committed 4 years ago

106b0b6a

[quant] dequantize support list and tuple of tensors (#41079)

jerryzh168 committed 4 years ago

edcf2cdf

Add fp16 support to SparseLengthSum PyTorch operator (#41058)

brad-mengchi committed 4 years ago

c8641584

[caffe2] Revert D22220798 (#41302)

Hao Lu committed 4 years ago

28291d3c

fix the range of the random weights used in the int8fc test (#41303)

Hector Yuen committed 4 years ago

e544bf29

Revert D22467871: add check for duplicated op registration in JIT

linbinyu committed 4 years ago

a1ed6e1e

[caffe2] Fix the issues when using CUB RadixSort (#41299)

jianyuh committed 4 years ago

095886fa

Solve log2(x:int) ambiguity by using log2(float(x)) (#41295)

malfet committed 4 years ago

d1f06da9

Fix arg type annotations in jit.trace and onnx.export (#41093)

JBamberger committed 4 years ago

1c098ae3

Ampere has CUDA_MAX_THREADS_PER_SM == 2048 (#41138)

zasdfgbnm committed 4 years ago

877a5996

Better THGeneric.h generation rules in bazel (#41285)

malfet committed 4 years ago

6cbb9249

Revert D22465221: [pytorch][PR] Reducing size of docker Linux image

malfet committed 4 years ago

67f5d68f

Define PSIMD_SOURCE_DIR when including FP16 (#41233)

Flamefire committed 4 years ago

ac3542fa

msvc anonymous namespace bug (#41199)

EscapeZero committed 4 years ago

abea7cd5

Disable the mkldnn for conv2d in some special cases (#40610)

lly-zero-one committed 4 years ago

48d6e2ad

[JIT] Add support for backend-lowered submodules (#41146)

Meghan Lele committed 4 years ago

ce3ba3b9

Impilcit casting resulting internal build failure. (#41272)

kimishpatel committed 4 years ago

1f2e91fa

Revert D22329069: Self binning histogram

malfet committed 4 years ago

7bae5780

[ONNX]Add tests for ConvTranspose 1D and 3D (#40703)

KsenijaS committed 4 years ago

dd0c98d8

Change to.dtype_layout to c10-full (#41169)

smessmer committed 4 years ago

9daba76b

Reducing size of docker Linux image (#41207)

Steve Sylvain committed 4 years ago

7c143e5d

Improve repr for torch.iinfo & torch.finfo (#40488)

Kiyosora committed 4 years ago

0651887e

Migrate addmm, addbmm and THBlas_gemm to ATen (#40927)

peterbell10 committed 4 years ago

cb6c3526

Self binning histogram (#40875)

Yavuz Yetim committed 4 years ago

16c8146d

[ONNX]Fix export of flatten (#40418)

KsenijaS committed 4 years ago

9b0393fc

add check for duplicated op registration in JIT (#41214)

linbinyu committed 4 years ago

a548c6b1

Wrap Caffe2's SparseLengthsSum into a PyTorch op (#39596)

nqjade committed 4 years ago

75b6dd3d

Small clarification of torch.cuda.amp multi-model example (#41203)

definitelynotmcarilli committed 4 years ago

d927aee3

LogitOp LUT based fake FP16 Op. (#41258)

venkatacrc committed 4 years ago

4a09501f

Newer Older