strisunshinewentingwang/pytorch

Pull Requests Commits

identify important circleci builds (#20498)

suo committed 5 years ago

24cd0e08

Remove dependencies from Caffe2Go on PyTorch JIT (#20463)

smessmer committed 5 years ago

9e7f22b2

UpSample GPU Porting (#19630)

xmnlab committed 5 years ago

34797775

Add ShapeInference for AtomicIter Op (#20021)

Cheng Cheng committed 5 years ago

7ffc37e0

Split nn.MultiHeadAttention into Module + functional (#20415)

jamarshon committed 5 years ago

6e82b1c7

Update Sleef to include fix for FMA4 detection (#20450)

colesbury committed 5 years ago

b46a6308

eliminate FE_INVALID exceptions related to fp16 conversion (#20390)

jspark1105 committed 5 years ago

10117687

codemode change missing [from D13586737]

Du Tran committed 5 years ago

8e9692df

Bump torch proto version (#20444)

davidriazati committed 5 years ago

e8fb5f35

add c2 benchmark runs in cpp (#20108)

ajyu committed 5 years ago

a9aaf698

temporarily disbale layernorm AD (#20442)

wanchaol committed 5 years ago

d2da3ee6

Rename AT_ASSERT to TORCH_INTERNAL_ASSERT; other macro updates (#20321)

ezyang committed 5 years ago

f0829f37

Fix version counter sharing in set_data() (#20391)

Will Feng committed 5 years ago

13641040

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

Newer Older