strisunshinewentingwang/pytorch

Pull Requests Commits

patching 11.1 ptxas issue

jjsjann123 committed 2 years ago

e48b29b1

functionalization bugfix: using owning type when unwrapping tensors

bdhirsh committed 2 years ago

640ce6bc

[NVFuser] multithreading nvfuser test

davidberard98 committed 2 years ago

f36d348f

Avoid compiling ts backend binidngs in fbcode build

wconstab committed 2 years ago

a4ce1136

[POC] fix static init issue with JIT container types

bdhirsh committed 2 years ago

40f3e850

remove _is_foreach_op codegen special cases, clean up mutable return type checks

bdhirsh committed 2 years ago

74e93f72

functionalization: add native fill() op

bdhirsh committed 2 years ago

ea5209c9

functionalization: add a copy() native function

bdhirsh committed 2 years ago

5da76acd

functionalization: add support for zero_()

bdhirsh committed 2 years ago

7d44b367

[complex32] mul

kshitij12345 committed 2 years ago

c8fd1bbc

[Pytorch][4/4 Static dispatch] Support multiple backends with multiple kernels (#76059)

priyaramani committed 2 years ago

f954c0a7

[metal] Fix error reporting on failure to flush command buffer (#76263)

SS-JIA committed 2 years ago

1df2d6a9

third_party: Fix build_bundled script

seemethere committed 2 years ago

a9a17ce6

[NVFuser] always use fallback if fusion fails

davidberard98 committed 2 years ago

da984c50

Distribute torchgen as part of PyTorch package

ezyang committed 2 years ago

5109d81f

Revert "[nnc] Strides to Tensor (#72962)"

zengk95 committed 2 years ago

1d555181

[flatbuffer] Bugfix: some class dont have __getstate__ (#76197)

qihqi committed 2 years ago

c55b425d

[FSDP] Support initialization of modules on meta device

rohan-varma committed 2 years ago

b4b1b0d5

Make sure requires_grad is propagated for all backend

albanD committed 2 years ago

e4d5801e

Re-land D35572928: [kineto] implement kineto client interface (#76239)

chaekit committed 2 years ago

920876d6

[Vulkan] VK Timestamp Queries for op profiling (#75829)

beback4u committed 2 years ago

4f3dd80c

move Bazel version header generation to shared build structure (#75332)

mikey dagitses committed 2 years ago

f4200600

[cmake] Use list(APPEND instead of string(APPEND for vulkan codegen args

smspillaz committed 2 years ago

111b2bf9

delete ${GEN_VULKAN_FLAGS}

MambaWong committed 2 years ago

bc9bba9b

add comment for lerp cuda implementation

jjsjann123 committed 2 years ago

7a27accb

define the caffe2_serialize target in Bazel (#75942)

mikey dagitses committed 2 years ago

d78dd825

[JIT] Add Context Manager to force strict fusion

Elias Ellison committed 2 years ago

0d7be81c

[EASY] [JIT]Cleaning up the device analysis temp test code

Gamrix committed 2 years ago

fdaf393e

Revert "[PyTorch] Add native fast path for transformer encoder inference"

bigfootjon committed 2 years ago

2387efd3

remove bucket_size_limit property from bucket struct

zhaojuanmao committed 2 years ago

1c35f37c

Newer Older