strisunshinewentingwang/pytorch

Pull Requests Commits

Improve hip-clang support in build_amd.py (#23835)

iotamudelta committed 5 years ago

fc368425

Fix test TestCuda.test_streams_multi_gpu_query (#23912)

yxsamliu committed 5 years ago

13a684d5

Update CosineAnnealingWarmRestarts to follow PyTorch 1.1+ Step Order. (#23833)

gchanan committed 5 years ago

fc82ec29

Change fbgemm_linear_{int8,fp16}_weight to fbgemm_linear_{int8,fp16}_weight_fp32_activation (#22955)

jianyuh committed 5 years ago

78cc9b92

Erase shape information from class types (#23362)

davidriazati committed 5 years ago

002d4f9f

Recursive script migration guide

davidriazati committed 5 years ago

b0a27278

sync and async torch.distributed.rpc for builtin operators (#23228)

mrshenli committed 5 years ago

8b349073

Set caffe2_tvm_min_ops to 8 (#23893)

Hao Lu committed 5 years ago

c07fc96b

Updating submodules

svcscm committed 5 years ago

ddc25efc

Rename cpu-only to cpuonly, as dash features are not supported. (#23879)

ezyang committed 5 years ago

68318404

Enable OSS quantization tests (#23858)

James Reed committed 5 years ago

40f0b1c8

Add flag to temporarily disable MKL-DNN conv (#23837)

James Reed committed 5 years ago

6ba60ec9

weight_names bug fix (#23848)

amylittleyang committed 5 years ago

9588cd92

format init.cpp (#23840)

Elias Ellison committed 5 years ago

d413f2d3

Updating submodules

svcscm committed 5 years ago

43c4bcba

Docs: Delete placeholder to use top-level file (#23869)

hugovk committed 5 years ago

52be1448

Fix SliceGradientOp to handle properly empty batches (#23784)

kennyhorror committed 5 years ago

d58059bc

Delete Travis CI config (#23788)

ezyang committed 5 years ago

c002ede1

Define toIValue conversion for dtype (#23708)

James Reed committed 5 years ago

489cc466

Rename previously THNN conv kernels to have naive_ prefix.

ezyang committed 5 years ago

e8cf9b68

Remove nightly suffix from nightlies; upload to pytorch-nightly. (#23752)

ezyang committed 5 years ago

60a4ef30

Properly mangle `nn.Module.__construct` (#23779)

suo committed 5 years ago

e2f5bc5c

fix some compiler warnings

suo committed 5 years ago

8fc349f7

fix torch.frac documentation (#23830)

Ailing Zhang committed 5 years ago

51d59a43

make nn.LSTM accept PackedSequence instead of Tuples

wanchaol committed 5 years ago

8fb0d198

Negate halves on GPU using __hneg() when possible, instead of using float conversion.

xuhdev committed 5 years ago

a1584555

Recommend `~` and `bitwise_not()` when user tries to apply neg (`-`) on a bool tensor.

xuhdev committed 5 years ago

f90afff3

Std opset export (#22310)

neginraoof committed 5 years ago

f278aee7

Add CUDA 10.1 to CI. (#23791)

ezyang committed 5 years ago

a710f816

galv committed 5 years ago

0015b188

Newer Older