strisunshinewentingwang/pytorch

Pull Requests Commits

Profiler: add Self CPU Time Total, CPU time total and other general improvements (#19378)

salexspb committed 5 years ago

7a8bc85f

Quantized SumRelu (#19319)

Zafar Takhirov committed 5 years ago

6e06154c

Use the QTensor with QReLU (#19312)

Zafar Takhirov committed 5 years ago

76307667

Changing the rounding in the QTensor (#19714)

Zafar Takhirov committed 5 years ago

db900849

Fix env vars needed for devtoolset7 binaries

pjh5 committed 5 years ago

e814c110

Fixing update_s3_htmls for binaries

pjh5 committed 5 years ago

c5cca653

Fix case for `activations` attribute in nn.RNN ONNX export. (#19368)

Spandan Tiwari committed 5 years ago

9ef8eb4c

Remove duplicate inlineCallToCode (#19724)

zdevito committed 5 years ago

a425e1cb

Serialize first-class version of functions (#19723)

zdevito committed 5 years ago

330990d8

Trace directly into first-class module form. (#19722)

zdevito committed 5 years ago

6cb1b994

@torch.jit.script(fn) now is a torch.jit.Function (#19721)

zdevito committed 5 years ago

31524bda

pybind CompilationUnit and Function directly (#19720)

zdevito committed 5 years ago

12f7c2de

caffe2 | Use _aligned_free in WorkerPool destruction (#19751)

Oleg Bogdanov committed 5 years ago

bf5a5c2a

Bug fix in bound shape inferencer (#19729)

Yinghai Lu committed 5 years ago

65496e4e

Fall back to asking nvcc for detecting cuda version if no *cudaart* is found (#19741)

t-vi committed 5 years ago

556c8a30

Automatic update of fbcode/onnx to 27d4b617e7097cda7d0d4c45ff2b09d248f33179 (#19718)

houseroad committed 5 years ago

5025d1d5

Fix Circle CI for ONNX repo (#19725)

houseroad committed 5 years ago

bbedaddd

Make interpolate bicubic match opencv result (#19703)

Ailing Zhang committed 5 years ago

0effe1d4

Fix compilation on Windows 10 (CUDA 10.0, Visual Studio 2017) (#19615)

dthul committed 5 years ago

29d8711e

Add SGDR(Stochastic Gradient Descent with Warm Restarts) scheduler (#17226)

Kirayue committed 5 years ago

af06d634

Replace cpu_apply with TensorIterator inside of Copy function (#18618)

VitalyFedyunin committed 5 years ago

465799fa

fix typo: pytoch -> pytorch

seungwonpark committed 5 years ago

6c7135de

try to make at::cat in mm_tree_reduction operate on contig tensors (#18816)

Natalia Gimelshein committed 5 years ago

3875e1ba

Fix the insert_guard for norm decomposation (#19646)

wanchaol committed 5 years ago

c5719691

add max_pool2d to AD, add tests for both autodiff and inference mode

wanchaol committed 5 years ago

3c81eb3a

Fix the depthwise 3x3x3 fast path criteria for the stride (#19692)

csummersea committed 5 years ago

cbd0a2d3

use relative path to load libthnvrtc (#19690)

Natalia Gimelshein committed 5 years ago

614871d9

Change some comments related to moving copy_ to native (#19618)

Roy Li committed 5 years ago

72b8b6c3

Remove old complex Types (#19616)

Roy Li committed 5 years ago

17e4cd0c

Remove function variant of copy_ (#19622)

Roy Li committed 5 years ago

6fead42e

Newer Older