strisunshinewentingwang/pytorch

Pull Requests Commits

Add annotations to PRs from forks (#54779)

samestep committed 3 years ago

56f12e61

Use custom sqrt if stdc++ does not fall back to C99 csqrt (#54820)

malfet committed 3 years ago

68af6d95

(BE) Refactor get-test-times-from-S3 into s3_stat_parser (#54808)

janeyx99 committed 3 years ago

717e70a8

Raise error in clip_grad_norm_ if norm is non-finite (#53843)

kurtamohler committed 3 years ago

3ddc6174

Restore storage on meta tensors; increase meta coverage (#53973)

ezyang committed 3 years ago

1f36ce6e

Adds the cfloat dtype to the eager and jit variant consistency tests (#54854)

Mike Ruberry committed 3 years ago

94efb48e

make all arguments in test_history.py optional kwarg (#54797)

Rong Rong (AI Infra) committed 3 years ago

2fd1eb3a

Using latest windows CUDA exe (#54596)

Rong Rong (AI Infra) committed 3 years ago

6d2bf76b

fix silent correctness bug with channels_last usage of upsample cuda kernels (#54744)

bdhirsh committed 3 years ago

86b1f4e9

Deprecate legacy constructor `torch.Tensor()` (#54414)

ysiraichi committed 3 years ago

4e5af53d

[quant][fx] store dtype, axis as literals in the graph (#54624)

supriyar committed 3 years ago

a0a7a2d6

Port torch.outer method_tests() to OpInfos (#54798)

lithuak committed 3 years ago

9e6877c5

[testing] support op with args/kwargs in test_unary_ufunc (#52194)

kshitij12345 committed 3 years ago

b7c5d575

enable bf16 for cat serial kernel (#54674)

haozhe.zhu committed 3 years ago

07350da3

enable bf16 vec copy (#54671)

haozhe.zhu committed 3 years ago

01b15570

[numpy] Add torch.take_along_dim (#52833)

kshitij12345 committed 3 years ago

0527d142

Make index_add take a scalar argument alpha (#54176)

zasdfgbnm committed 3 years ago

eec48303

optimizer exploration - v1 and v2 + fix position_weighted optimizer + decoupled weight decay (#54042)

lanlanfb committed 3 years ago

695eef05

[DDP] Mark a few variables as const in reducer (#54764)

rohan-varma committed 3 years ago

5c3d80d8

[c10d] s/torch::autograd::variable/at::Tensor/g (#54763)

rohan-varma committed 3 years ago

671f80a3

Implement public API InferenceMode and its error handling (#53343)

Ailing Zhang committed 3 years ago

7caa4646

Compute Tensor::toString() without reference to backend (#54711)

ezyang committed 3 years ago

23091731

Make memory overlap a little less precise so it works with null data ptr (#54710)

ezyang committed 3 years ago

f0679725

Make the fuser raise NotImplementedError when unknown device is hit (#54709)

ezyang committed 3 years ago

c782949e

Avoid testing device in cdist when called in a "Math" context (#54708)

ezyang committed 3 years ago

6445c9a1

Make convolution_overrideable default implementation raise NotImplementedError (#54707)

ezyang committed 3 years ago

c9e0aab2

Disambiguate where 'Doesn't run' error message comes from (#54706)

ezyang committed 3 years ago

ed560cf2

Fix missing format string qualifier (#54705)

ezyang committed 3 years ago

b5ab3482

Rename linalg.det test so that it generates a valid method name (#54704)

ezyang committed 3 years ago

d9a7c758

Add empty_generic, which allocates an empty tensor in a device-generic way (#54703)

ezyang committed 3 years ago

05fa570b

Newer Older