strisunshinewentingwang/pytorch

Pull Requests Commits

Add device_check place holder for functions (#56870)

wenleix committed 3 years ago

183320df

Fix tensor device in test_kthvalue_overlap (#56869)

wenleix committed 3 years ago

f7f85407

[PyTorch] Autoformat c10 (#56830)

swolchok committed 3 years ago

44cc873f

[pytorch][nnc] update external functions for mobile build (#56850)

ljk53 committed 3 years ago

3c4d57c1

[PyTorch] Take advantage of string literals in TORCH_WARN (#54032)

swolchok committed 3 years ago

b11a2420

[RPC Framework] Enable RemoteModule to directly send GPU tensors over the wire on TensorPipe RPC backend if a device map is provided (#57288)

Yi Wang committed 3 years ago

13dbb77b

[torch][segment_reduce] Add support for initial value (#56923)

Serhat Yilmaz committed 3 years ago

20eac093

Added sm_75 support for CI Xenial CUDA 11.1 cuDNN 8 builds (#57320)

mstfbl committed 3 years ago

bd347012

Clean up naming and comments (#56964)

soulitzer committed 3 years ago

2b54cec7

Refactor fast gradcheck (#55871)

soulitzer committed 3 years ago

bbdadab3

[nnc] ported some more ops + added vectors to argvalue (#56766)

Chillee committed 3 years ago

47e9ec40

Maintain submodule references during subgraph rewriting (#55463)

Ansley Ussery committed 3 years ago

233f2cd2

[pytorch] fewer cuda sync in unique by using cub instead of thrust (#57323)

jspark1105 committed 3 years ago

3a5f8546

[PyTorch] ifdef out ATen tests that fail with static dispatch (#57379)

swolchok committed 3 years ago

208f81b7

Fix min() and max() for empty tensors (#52565)

v0dro committed 3 years ago

293830bc

[JIT] Enable conv-add-relu fusion as a part of frozen graph optimization (#56580)

desertfire committed 3 years ago

c1a44224

[StaticRuntime] Use NNC's call_raw API to reduce call overheads. (#57329)

Mikhail Zolotukhin committed 3 years ago

400ca767

[TensorExpr] Add `TensorExprKernel::runFast` method. (#57328)

Mikhail Zolotukhin committed 3 years ago

f219ed66

[TensorExpr] Add `CodeGen::call_raw` method. (#55113)

Mikhail Zolotukhin committed 3 years ago

c9ab384a

Fix binary_checkout to use master (#57389)

davidriazati@fb.com committed 3 years ago

4c3283da

per_channel fake quant fp16 and fp64 support (#56894)

HDCharles committed 3 years ago

5e422fa1

Fix NVRTC versioning for CUDA 11.X (X>=3), CUDA 12 and later (#57204)

zasdfgbnm committed 3 years ago

42b3fc29

[8/n] [torch/elastic] Add unit tests for _RendezvousState (#56536)

cbalioglu committed 3 years ago

72b1faa2

[CUDA graphs] [BC-breaking] Makes torch.cuda.amp.GradScaler scale updates in-place for better composability with graph capture (#55562)

mcarilli committed 3 years ago

bbc3cc67

[PyTorch] Optimize intrusive_ptr(TTarget*) ctor (pybind) (#57053)

swolchok committed 3 years ago

3a777b67

Revert D28011862: Add pybind interface to caffe2 quantization server

samestep committed 3 years ago

b9b768c0

Fix MAGMA qr for empty batched inputs (#56257)

IvanYashchuk committed 3 years ago

f54aa85a

Add cuSOLVER path for torch.linalg.qr (#56256)

IvanYashchuk committed 3 years ago

ff59039a

Remove size arguments for internal orgqr and geqrf calls (#56255)

IvanYashchuk committed 3 years ago

6cb9abfd

Add non-allocating helper function for torch.linalg.qr (#56254)

IvanYashchuk committed 3 years ago

d5e1cac6

Newer Older