strisunshinewentingwang/pytorch

Pull Requests Commits

C++ API parity: MaxUnpool3d

pbelevich committed 4 years ago

5005f7bc

Fix ONNX Interpolate

lara-hdr committed 4 years ago

f4f6d8dd

make python udf serialization format to be binary plus tensor tables (#27136)

zhaojuanmao committed 4 years ago

631e2ee7

C++ API parity: MaxUnpool2d

pbelevich committed 4 years ago

5cac7387

Makes more of test_cuda.py generic and updates test_torch tests (#27135)

Mike Ruberry committed 4 years ago

b45f1b96

Add some extra info to PyTorchStreamReader/Writer errors (#27006)

Basil Hosmer committed 4 years ago

2dff0b6f

add dynamic isinstance (#26269)

zdevito committed 4 years ago

becf080e

Allow types as node attributes (#26268)

zdevito committed 4 years ago

8a38a53e

Add test case for init_rpc_backend (#26997)

xush6528 committed 4 years ago

00e58829

Serialize XLA Tensor (#27041)

Ailing Zhang committed 4 years ago

0b79f77a

Revert D17694691: Enable distributed autograd tests for >py36

mrshenli committed 4 years ago

bb8983e9

Handle uninitialized min/max values in histogram observer (#27151)

raghuramank100 committed 4 years ago

4abfb549

C++ API parity: MaxUnpool1d

pbelevich committed 4 years ago

d125a83f

Enable distributed autograd tests for >py36

mrshenli committed 4 years ago

7bbb2df6

Unify quantized conv and linear tests (#26992)

supriyar committed 4 years ago

b805b5da

Updating submodules

svcscm committed 4 years ago

38001fec

Move comment back to the class it describes (it accidentally drifted away).

Mikhail Zolotukhin committed 4 years ago

8e4031dd

Disable flaky distributed autograd test under spawn mode

mrshenli committed 4 years ago

e09c97b1

Enable tests (#27103)

jerryzh168 committed 4 years ago

98c02e6d

insert_prepack_unpack pattern (#27102)

jerryzh168 committed 4 years ago

14d29aee

API - add more passes to graph mode (#27093)

jerryzh168 committed 4 years ago

f742ceaa

Uninitialize the accumulation buffer to save some overhead (#27005)

jianyuh committed 4 years ago

6a096764

make rpc and dist-autograd multiprocess test to use both fork and spawn (#25656)

zhaojuanmao committed 4 years ago

1d2d59dd

Ensure that DDP wrapped module has parameters that require gradients (#25858)

danthe3rd committed 4 years ago

46539eee

Fixed seek offset size to 64bit. (#27125)

peterjc123 committed 4 years ago

ec07d144

Delete 5 THVector function declarations that have already been migrated to ATen.

xuhdev committed 4 years ago

48cd66cd

Bump gloo (#27133)

pietern committed 4 years ago

4086f743

Add C++ torch::nn::functional::pdist (#27122)

jon-tow committed 4 years ago

18eea826

Automatically select proper tqdm submodule (#27108)

ssnl committed 4 years ago

5ccecabb

Creates device generic cuDNN decorators (#26791)

Mike Ruberry committed 4 years ago

30997320

Newer Older