strisunshinewentingwang/pytorch

Pull Requests Commits

fake_quant: add a more memory efficient version

vkuzo committed 4 years ago

8de18ba0

Fix replication_pad for cuda launch configuration (#50565)

xwang233 committed 4 years ago

db86dd8a

Fix indexing for overrides. (#49324)

hameerabbasi committed 4 years ago

cf1882ad

[ROCm] re-enable tests (#50691)

KyleCZH committed 4 years ago

16faabe7

Update JIT_OPT macro for easier use (#50602)

tugsbayasgalan committed 4 years ago

fbf7eec8

Enable TensorPipe's CMA channel (#50759)

lw committed 4 years ago

112a5834

[metal] Use MPSCNN kernels for binary elementwise ops

Stephen Jia committed 4 years ago

c18403a6

[PyTorch] Remove CAFFE2_FB_LIMITED_MOBILE_CAPABILITY (#50385)

swolchok committed 4 years ago

1e0809db

Fix test_dispatch.py when running with TORCH_SHOW_CPP_STACKTRACES=1 (#50509)

zasdfgbnm committed 4 years ago

4f3cdd97

JIT Testing: Improve assertAutodiffNode error message (#50626)

Lillian Johnson committed 4 years ago

f1c57859

Add complex autograd support and OpInfo based test for torch.addr (#50667)

anjali411 committed 4 years ago

1cc8f8a7

tools: Move sha check to else statement (#50773)

seemethere committed 4 years ago

66adfcd2

Issue #48724. Only set the CMake IMPORTED_LOCATION property in static… (#49173)

foundry-markf committed 4 years ago

e1bb4769

[WIP] JIT Static Hooks: cpp tests (#49547)

Lillian Johnson committed 4 years ago

22902b92

[WIP] JIT Static Hooks: python tests (#49546)

Lillian Johnson committed 4 years ago

3b88e1b0

[WIP] JIT Static Hooks: serialization logic (#49545)

Lillian Johnson committed 4 years ago

0eb41e67

[WIP] JIT Static Hooks: schema checking logic (#49975)

Lillian Johnson committed 4 years ago

9c494572

[WIP] JIT Static Hooks: adding hooks to class type and adding logic for hook running/compilation (#49544)

Lillian Johnson committed 4 years ago

a722d28e

Revert D25958987: [pytorch][PR] Add type annotations to torch.overrides

mrshenli committed 4 years ago

1f5c3b3a

Add new backend type for Intel heterogeneous computation platform. (#49786)

chengjunlu committed 4 years ago

4a8ef452

Let TensorPipe detect peer access (#50676)

mrshenli committed 4 years ago

a3b8cbcd

Implement NumPy-like function torch.fmax() & torch.fmin() (#49312)

kiyosora committed 4 years ago

4803eaf5

Add type annotations to torch.overrides (#48493)

guilhermeleobas committed 4 years ago

2ace4fc0

[JIT] Fix archive file extension in examples and docs (#50649)

Meghan Lele committed 4 years ago

4aea0073

[quant] Add non-fbgemm fallback implementation for embedding lookup ops (#50706)

supriyar committed 4 years ago

e0096650

[FX] Fix NoneType annotation in generated code (#50777)

James Reed committed 4 years ago

5205cc1c

[JIT] Print out CU address in `ClassType::repr_str()` (#50194)

Meghan Lele committed 4 years ago

8f5ad00e

Cat benchmark: use mobile feed tensor shapes and torch.cat out-variant (#50778)

Marat Subkhankulov committed 4 years ago

dea9af5c

Generalize `sum_intlist` and `prod_intlist`, clean up dimensionality functions (#50495)

r-barnes committed 4 years ago

06c734d8

Fix Native signature for optional Tensor arguments (#50767)

smessmer committed 4 years ago

47c57b88

Older