strisunshinewentingwang/pytorch

Pull Requests Commits

Guard {set,rebase}_history on grad_fn check (#19623)

James Reed committed 5 years ago

80020b3d

optimize BatchMatmulOp (#18612)

xiaomengy committed 5 years ago

fb9fc42a

fix lint (#19632)

jerryzh168 committed 5 years ago

176bdc07

Add base support to torch.logspace, default base=10 (#19542)

Phúc Lê committed 5 years ago

9b272aff

disable flake8 E302 (two blank lines) (#19634)

suo committed 5 years ago

96b96629

fix nn.Sequential doc

ssnl committed 5 years ago

6b8771a7

caffe2 | Windows compat fixes

Oleg Bogdanov committed 5 years ago

70b82d28

Remove fixed TODO (#19590)

smessmer committed 5 years ago

2e048feb

correct comments in group_norm_op (#19621)

hl475 committed 5 years ago

55e53d3d

Simplify argument test cases (#19593)

smessmer committed 5 years ago

5f82d59c

Add test cases for optional of list (#19592)

smessmer committed 5 years ago

fddd763e

Port adaptive_max_pool3d() to ATen (#19547)

skrah committed 5 years ago

fc8834df

add torch.tensor requires grad (#19445)

Elias Ellison committed 5 years ago

0922a64d

Surface the Glow traces to C2 (#19087)

Yinghai Lu committed 5 years ago

4e8cc8ee

Fix lack of state init for adagrad and add share_memory flag (#17679)

Stonesjtu committed 5 years ago

444f792f

Allow extracting element-wise loss in softmax (#19579)

prigoyal committed 5 years ago

0d0acba3

dispatch max_pools with no indices, expose max_pools to torch namespace (#19449)

wanchaol committed 5 years ago

e9c8f372

Adds `fakeQuantizePerTensorAffineOp` to pytorch (#19387)

jerryzh168 committed 5 years ago

f3be2816

-fno-math-errno -fno-trapping-math (#19552)

James Reed committed 5 years ago

1b3967b4

Only require python print on certain namespaces (#19383)

bwasti committed 5 years ago

d8729efa

Use `fbgemm` for quantize/dequantize ops (#19500)

Zafar Takhirov committed 5 years ago

3cc60e54

Specify to use Float16UniformFill if necessary in sparse lookup layer (#18499)

chocjy committed 5 years ago

714344a9

Fix the Division by Zero Bug of CosineAnnealingLR (#19180)

chandlerzuo committed 5 years ago

e3f15046

Fix the documentation for BCEWithLogitsLoss (#17218, #16804) (#19212)

vvelicodnii-sc committed 5 years ago

7a418969

fix the docstring of `RandomSampler` (#19113)

crcrpar committed 5 years ago

bb05f707

Avoid (future) cusparse name collision (#19591)

mruberry committed 5 years ago

83cf9473

Add docs and test guaranteeing indices from torch.nonzero ordered C-style (#19539)

jhultman committed 5 years ago

f767c9ac

Remove unnecessary printing from tests

ssnl committed 5 years ago

3b4d4ef5

Fix lr_scheduler's last_epoch value at the time of initialization (BC BREAKING!) (#7889)

bado-lee committed 5 years ago

36084908

Removes variable which is assigned but not used (#19194)

SebFar committed 5 years ago

f9c4ce78

Newer Older