strisunshinewentingwang/pytorch

Pull Requests Commits

MIOpen integration into pytorch RNN operators (#22774)

ashish committed 5 years ago

9f8e2c06

Speeds up fast-path for 1D tensors (#22756)

colesbury committed 5 years ago

30e03df6

avoid kernel launches for zero-sized tensor inputs

wanchaol committed 5 years ago

02bc06a6

Make PythonArgs::tensor and PythonArgs::scalar faster (#22782)

colesbury committed 5 years ago

b1b65f34

optimize topk on cpu using parallel and partial sort (#19736)

mingfeima committed 5 years ago

10c14ad1

Speed up TensorIterator::compute_strides a little (#22779)

colesbury committed 5 years ago

fc23d7f3

Initiate checkCuda90Bug warning (#22757)

Guanheng Zhang committed 5 years ago

f266a63e

Revert D16222539: [pytorch][PR] Let users pass CMake-specific options starting with CMAKE_ to CMake.

ezyang committed 5 years ago

ccb28939

Let users pass CMake-specific options starting with CMAKE_ to CMake. (#22776)

xuhdev committed 5 years ago

612eed31

add new tests to benchmark_all_test (#22787)

mingzhe09088 committed 5 years ago

7eb03193

make custom op work in OSS environment (#22781)

mingzhe09088 committed 5 years ago

1878800f

Add the support of handle Bias being nullptr for torch.ops.quantized.fbgemm_linear (#22403)

jianyuh committed 5 years ago

8ec712da

Update ONNX constant folding to support opset 10. (#22515)

Spandan Tiwari committed 5 years ago

9d11004e

make CompilationUnit::define return defined functions

suo committed 5 years ago

291570e0

refactor self to be a class again

suo committed 5 years ago

de819be9

Give functions qualified names

suo committed 5 years ago

22d70e0d

Suppress progress bar only for pip install

ezyang committed 5 years ago

4b48ae4a

Remove hard-coded NVRTC specific constant from fuser header

bwasti committed 5 years ago

05d56bd1

assert_no_internal_overlap pass op name by const ref

James Reed committed 5 years ago

513b7a7a

Move the storage in empty_cpu

James Reed committed 5 years ago

9690f862

bucketize op shape inference (#22716)

ChunliF committed 5 years ago

a7978151

Back out "[pytorch][PR] Move thnvrtc and DynamicLibrary to ATen" (#22749)

ezyang committed 5 years ago

ac78a86e

optimize RNN on CPU (#22512)

xiaomengy committed 5 years ago

8bdda03a

jjsjann123 committed 5 years ago

3135298d

Improve hypothesis_utils.py for qtensor (#22693)

jerryzh168 committed 5 years ago

1682d38a

bddppq committed 5 years ago

3fabb9f1

add fill_diagonal function (#21892)

shihongzhi committed 5 years ago

45cf33a7

Add environment variables used in CONTRIBUTING example (#22736)

kianasun committed 5 years ago

89d6e880

enabled MIOpen depthwise convolutions (#22696)

lcskrishna committed 5 years ago

5147819f

Quantized Conv2d Module (#21323)

z-a-f committed 5 years ago

d21e476d

Newer Older