strisunshinewentingwang/pytorch

Pull Requests Commits

fix quantization mechanism to match nnpi (#41494)

Hector Yuen committed 4 years ago

26790fb2

resurrect single quantization op test (#41476)

Hector Yuen committed 4 years ago

e6859ec7

enable TE on windows (#41501)

Krovatkin committed 4 years ago

04c0f2e3

Rename capacity to nbytes in ShareExternalPointer to avoid confusion in future (#41461)

houseroad committed 4 years ago

b2e52186

Move GLOG_ constants into c10 namespace (#41504)

malfet committed 4 years ago

70214075

[Gloo] update gloo submodule for PyTorch (#41462)

jiayisuse committed 4 years ago

f27e395a

onnx export of fake quantize functions (#39738)

skyw committed 4 years ago

1fb2a7e5

[PyTorch Mobile] Modularize the autograd source files shared by mobile and full-jit (#41430)

iseeyuan committed 4 years ago

7a33d8b0

[reland] Enable TF32 support for cuBLAS (#41498)

zasdfgbnm committed 4 years ago

23174ca7

Implement gcd, lcm (#40651)

aayn committed 4 years ago

200c3431

[jit] Fix jit not round to even if const is folded (#40897)

cloudhan committed 4 years ago

e44f4600

Restore the contiguity preprocessing of linspace (#41286)

xuhdev committed 4 years ago

1770937c

remove use of the term "blacklist" from docs/cpp/source/Doxyfile (#41450)

morrme committed 4 years ago

d90fb72b

Disable failed caffe2 tests for BoundShapeInference on Windows (#41472)

peterjc123 committed 4 years ago

404799d4

Updates serialization note to explain versioned symbols and dynamic versioning (#41395)

Mike Ruberry committed 4 years ago

60f2fa6a

Support @torch.jit.unused on a @torch.no_grad decorated function (#41496)

ppwwyyxx committed 4 years ago

488ee379

Reduce Image Size (2) (#41301)

Steve Sylvain committed 4 years ago

71c3b397

remove blacklist reference (#41447)

morrme committed 4 years ago

5bd71259

Learnable Fake Quantizer Benchmark Test (#41429)

Paul Shao committed 4 years ago

b7147fe6

[reland][DNNL]:enable batchnorm3d (#40995)

XiaobingSuper committed 4 years ago

2b8db35c

[reland][DNNL]:enable conv3d (#40691)

XiaobingSuper committed 4 years ago

b48ee175

Add C++ end to end test for RPC and distributed autograd. (#36893)

pritamdamania committed 4 years ago

ff6e5603

Pull upstream select_compute_arch from cmake for Ampere (#41133)

zasdfgbnm committed 4 years ago

8940a4e6

Cuda Support for Learnable Fake Quantize Per Channel (GPU) (#41262)

Paul Shao committed 4 years ago

c62550e3

Cuda Support for Learnable Fake Quantize Per Tensor (GPU) (#41127)

Paul Shao committed 4 years ago

4367a733

Adding epsilon input argument to the Logit Op

venkatacrc committed 4 years ago

225289ab

Revert D22480638: [pytorch][PR] Add non-deterministic alert to CUDA operations that use `atomicAdd()`

mrshenli committed 4 years ago

954c2600

[quant][pyper] Add embedding_bag weight quantize and dequantize ops (#41293)

supriyar committed 4 years ago

008ab27b

DDP Communication Hook Main Structure (#40848)

sinannasir committed 4 years ago

d5ae4a07

[cmake] Use PROJECT_SOURCE_DIR instead of CMAKE_* (#41387)

powderluv committed 4 years ago

c86699d4

Newer Older