strisunshinewentingwang/pytorch

Pull Requests Commits

Move at::NonVariableTypeMode to TensorImpl, and check it in is_variable() (#20392)

Will Feng committed 5 years ago

3a0b27b7

Automatic update of fbcode/onnx to e08efaa35ed54362dfa283240506c003175889b7 (#20443)

houseroad committed 5 years ago

2dc9152d

Needed fixes for binaries

pjh5 committed 5 years ago

824d4f99

Make sure reducer=None is not used when fp16 embedding is enabled

chocjy committed 5 years ago

6c3b8a24

davidriazati committed 5 years ago

63c05bff

Port adaptive_avg_pool3d to ATen (#19898)

crcrpar committed 5 years ago

7799ea5e

Remove support for CUDA 8 (#20298)

syed-ahmed committed 5 years ago

5268b7df

Tiny spelling mistake fix. (#20425)

sebftw committed 5 years ago

62957ab0

Move THCTensor_(uniform) to ATen (#20292)

syed-ahmed committed 5 years ago

67414714

math module support: gcd, copysign, erf, erfc, expm1, fabs, gamma, lgamma (#19707)

eugenekoran committed 5 years ago

5f7ef09f

Disable incremental_state function in MultiheadAttention module. (#20177)

Guanheng Zhang committed 5 years ago

41673d47

Make a deep copy of extra_compile_flag dictionnary (#20221)

ClementPinard committed 5 years ago

f8aa6a8f

Hotfix for caffe2 windows build (#20417)

peterjc123 committed 5 years ago

30bdb8c0

DataLoader: add error detection for worker_init_fn (#20150)

ssnl committed 5 years ago

f496ea36

Fix bug in non_blocking copy (#20305)

Roy Li committed 5 years ago

163f0e18

Add quant-dequant nodes for weights

nishantpdce committed 5 years ago

6a8f5579

Profiling GraphExecutor

Krovatkin committed 5 years ago

9499c7b7

Support Exports to Multiple ONNX Opset (#19294)

lara-hdr committed 5 years ago

f4d9bfaa

move DistillBatchLRLoss Layer from open source to fb

Xue Feng committed 5 years ago

1129b334

make trace's errors more helpful in terms of what it can and can't do when tracing module's methods

Krovatkin committed 5 years ago

3f3ee560

Add multiline type annotation support for Python frontend (#14922)

davidriazati committed 5 years ago

7d3d5b73

Fix reflection on weak modules, copy attributes (#20190)

davidriazati committed 5 years ago

3a39ce0f

Add all list specializations to pickler (#20191)

davidriazati committed 5 years ago

00d0ddb1

Eliminate a const_cast.

Owen Anderson committed 5 years ago

6197eed4

quantize_rnn_modules in ensemble_export (#20365)

jhcross committed 5 years ago

a4ae6896

Preserve log_dir arg and member for SummaryWriter (#20382)

orionr committed 5 years ago

a0c28291

Add documentation to Dispatch.h (#20339)

ezyang committed 5 years ago

3aa414c8

Avoid unnecessary refcount bump in unary operators. (#20331)

ezyang committed 5 years ago

10a9ef83

profile on uses

Krovatkin committed 5 years ago

75c6d37b

Remove CPU_tensor_parallel_kernel_apply2 (#20207)

cpuhrsch committed 5 years ago

c6255a57

Newer Older