strisunshinewentingwang/pytorch

Pull Requests Commits

CheckpointSequential support non-reentrant (#86331)

rohan-varma committed 1 year ago

7a411952

Docs: fix typo (#86273)

cherrywoods committed 1 year ago

3037f3d7

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

pytorchmergebot committed 1 year ago

233d6f19

Fix for binary upload step, use bash shell rather then default sh (#86382)

atalman committed 1 year ago

bf746798

[ao] fixing public v private for qconfig.py (#86026)

HDCharles committed 1 year ago

facf210f

[kineto] guard global observer init against Edge profiler (#86347)

chaekit committed 1 year ago

7c5e07f8

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

jiaxuzhu92 committed 1 year ago

bc919ac7

Two small improvements to references (#86371)

lezcano committed 1 year ago

08780229

Add total GPU memory utilization (#86250)

huydhn committed 1 year ago

795906f2

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

ZainRizvi committed 1 year ago

1059d3b5

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

pearu committed 1 year ago

6b295cd0

Support autograd on sparse_mm in full. (#86301)

pearu committed 1 year ago

8f2c2167

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

pearu committed 1 year ago

88b882cd

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

pearu committed 1 year ago

f104490d

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

pearu committed 1 year ago

fc21cc82

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

pytorchmergebot committed 1 year ago

bed1ece9

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

KinglittleQ committed 1 year ago

eb32330d

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

huydhn committed 1 year ago

b8b564c9

Remove extra bracket, update header definition (#86317)

cpuhrsch committed 1 year ago

0c148a4b

Use FindCUDAToolkit to find cuda dependencies (#82695)

peterbell10 committed 1 year ago

fb9b9659

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

malfet committed 1 year ago

fa799132

Inline asIntArrayRef (#86350)

ezyang committed 1 year ago

4d772889

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

andrewor14 committed 1 year ago

cebf08af

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

Antoni Viros i Martin committed 1 year ago

cdbffa7f

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

jmdetloff committed 1 year ago

85c3b745

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

PaliC committed 1 year ago

936e9305

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

slgong-fb committed 1 year ago

27c3fb03

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

slgong-fb committed 1 year ago

a117fde8

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

albanD committed 1 year ago

fd5085c4

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

ezyang committed 1 year ago

0a75c42f

Newer Older