strisunshinewentingwang/pytorch

Pull Requests Commits

Update distributed.rst backend collective support chart

H-Huang committed 2 years ago

Verified 0042eae0

[torch.ao.quantization] include torch.qint32 for static quant (#86345)

jiaxuzhu92 committed 2 years ago

bc919ac7

Two small improvements to references (#86371)

lezcano committed 2 years ago

08780229

Add total GPU memory utilization (#86250)

huydhn committed 2 years ago

795906f2

Make mergebot message clearer when starting a new merge (#86311)

ZainRizvi committed 2 years ago

1059d3b5

Enable autograd on Linear with sparse COO weight (#86302)

pearu committed 2 years ago

6b295cd0

Support autograd on sparse_mm in full. (#86301)

pearu committed 2 years ago

8f2c2167

Support sum on a sparse COO tensor. (#86300)

pearu committed 2 years ago

88b882cd

Support autograd on Linear with sparse compressed weight. (#86137)

pearu committed 2 years ago

f104490d

Enable sparse_dim() and dense_dim() methods for Strided tensors (#86203)

pearu committed 2 years ago

fc21cc82

[torchdynamo hash update] update the pinned torchdynamo hash (#86306)

pytorchmergebot committed 2 years ago

bed1ece9

Fix memory leak in _LRScheduler.step() (#85602)

KinglittleQ committed 2 years ago

eb32330d

Ensure the minimum NVIDIA driver version to be 515.57 for CUDA 11.7 (#86344)

huydhn committed 2 years ago

b8b564c9

Remove extra bracket, update header definition (#86317)

cpuhrsch committed 2 years ago

0c148a4b

Use FindCUDAToolkit to find cuda dependencies (#82695)

peterbell10 committed 2 years ago

fb9b9659

[MPS] Better error message for `slow_conv2d_forward` (#86303)

malfet committed 2 years ago

fa799132

Inline asIntArrayRef (#86350)

ezyang committed 2 years ago

4d772889

[Quant] Remove weight from DTypeConfig for non-weighted ops (#86335)

andrewor14 committed 2 years ago

cebf08af

🦊 [AI Accelerators] Consolidate native_layer_norm for nested tensor (#86295)

Antoni Viros i Martin committed 2 years ago

cdbffa7f

Conditionally build the TestApp benchmark based on lite interpreter (#86314)

jmdetloff committed 2 years ago

85c3b745

Delete torch::deploy from pytorch core (#85953)

PaliC committed 2 years ago

936e9305

[Profiler] trace verbose=false by default (#86263)

slgong-fb committed 2 years ago

27c3fb03

[Profiler] Apply TensorMetadata for Optimizer and nnModule (#86047)

slgong-fb committed 2 years ago

a117fde8

Symintify getitem and add the required helper functions (#86207)

albanD committed 2 years ago

fd5085c4

Workaround MSVC ICE due to constexpr char* template argument (#86288)

ezyang committed 2 years ago

0a75c42f

Add at::symint:: namespace for ease of templated functions (#86329)

ezyang committed 2 years ago

45f03d69

Reduce warning suppression by just disabling pytest warnings plugin (#86255)

ezyang committed 2 years ago

ea21a982

Add option to record C++ backtraces in _record_memory_history (#86145)

ezyang committed 2 years ago

adf59197

Refactor _cuda_recordMemoryHistory to use pybind11 (#86139)

ezyang committed 2 years ago

97d6b5bb

Add Context Manager for Disabling Multithreading in Backwards, use in aot autograd (#86245)

eellison committed 2 years ago

d0488932

Older