strisunshinewentingwang/pytorch

Pull Requests Commits

Add interpolate-style overloads to aten::upsample* ops (#37176)

dreiss committed 4 years ago

7183fd20

Add support for float[]? arguments in native_functions.yaml (#37175)

dreiss committed 4 years ago

fb9e44f8

Back out "Revert D22329069: Self binning histogram" (#41313)

Yavuz Yetim committed 4 years ago

d04a2e4d

.cirlceci: Setup nvidia runtime for cu as well (#41268)

seemethere committed 4 years ago

86d803a9

reduce logging for layernorm (#41305)

Hector Yuen committed 4 years ago

dea39b59

Pass the number of indices but not embedding size in PyTorch operator (#41315)

brad-mengchi committed 4 years ago

67a4f375

Impl for ParameterList (#41259)

arktrail committed 4 years ago

98df9781

Fake Quantization Per Channel Kernel Core Implementation (CPU) (#41037)

Paul Shao committed 4 years ago

fa153184

Fake Quantization Per Tensor Kernel Core Implementation (CPU) (#41029)

Paul Shao committed 4 years ago

5e72ebed

[quant] Adding zero point type check for per channel quantization (#40811)

z-a-f committed 4 years ago

402be850

[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

Newer Older