strisunshinewentingwang/pytorch

Pull Requests Commits

[PyTorch][Static Runtime] Fast path for contiguous to_copy (#65499)

swolchok committed 2 years ago

ffede499

[PyTorch] Try removing Android strtod implementation (#65713)

swolchok committed 2 years ago

7b10a76e

[PyTorch] Fix many Tuple::elements() callsites (#64065)

swolchok committed 2 years ago

176d3c6f

[fx2trt]fix slice tensor converter (#65960)

842974287 committed 2 years ago

f14e5e63

[PyTorch][easy] Use copy-and-move instead of copy-and-swap in IValue::operator= (#65826)

swolchok committed 2 years ago

21eebc9f

[fx][const_fold] Refactor to use base split module to simplify, and correctly handle non-single-Tensor outputs (#65933)

jfix71 committed 2 years ago

592481a5

[iOS][CI] Update dev certs (#66004)

xta0 committed 2 years ago

34682377

Revert D31317680: [pytorch][PR] Avoid saving self for`softmax` and `log_softmax`

suo committed 2 years ago

ccf8d48f

suppress mypy error (#66003)

suo committed 2 years ago

21da6ae9

Revert "Port `sort` kernel to structured kernels. (#62391)" (#65876)

bdhirsh committed 2 years ago

eac218db

Avoid saving self for`softmax` and `log_softmax` (#65242)

soulitzer committed 2 years ago

5f7cadc7

Fix internal assert failure for torch.all and torch.any with requires_grad=True (#65714)

IvanYashchuk committed 2 years ago

383c0a38

Make autograd codegen for differentiable outputs safer to use (#65823)

IvanYashchuk committed 2 years ago

53c0d91d

[nnc] Add BufHandle.store to python API (#65213)

bertmaher committed 2 years ago

bff8d8fd

[nnc] Add call_with_numel interface for fast CUDA calls (#65213)

bertmaher committed 2 years ago

8cf047af

Avoid UB when indexing into size-0 tensors (#65878)

bertmaher committed 2 years ago

8595b6ee

Improve pytorch type hints (Dataloader, trig functions)

shapovalov committed 2 years ago

fc52f129

[Static Runtime] Fuse ListUnpack + gather_ranges_to_dense (#65116)

Mike Iovine committed 2 years ago

982ef883

pytorch quantization ao migration phase 2: caffe2/test (#65832)

vkuzo committed 2 years ago

227e37dd

pytorch quantization ao migration phase 2: torch/jit (#65829)

vkuzo committed 2 years ago

dac35b35

pytorch quantization ao migration phase 2: caffe2/benchmark (#65833)

vkuzo committed 2 years ago

e3af4be9

[special] special alias for softmax (#62251)

kshitij12345 committed 2 years ago

c1447f06

[sparsity] Add m-out-of-n support in the WeightNormSparsifier (#65295)

z-a-f committed 2 years ago

c27b427c

[sparsity] Change API to take FQNs as configuration (#65296)

z-a-f committed 2 years ago

8b1aa853

[PyTorch] Avoid string construction from const char* and speedup empty string creation if error messages are suppressed (#65939)

dhruvbird committed 3 years ago

ea0de37d

Added jit log stream changing function and some refactor (#65768)

narang99 committed 3 years ago

2828ce53

[deploy][1/n] Make deploy code conform to PyTorch style. (#65861)

suo committed 3 years ago

33c03cb6

[TensorExpr] Move lowerings registration from kernel.cpp to lowerings.cpp. (#65553)

Mikhail Zolotukhin committed 3 years ago

765b6a90

[TensorExpr] Move 'compute*' functions to operators/... (#65552)

Mikhail Zolotukhin committed 3 years ago

015e0079

[TensorExpr] Port NNC lowerings to the new registry mechanism. (#65551)

Mikhail Zolotukhin committed 3 years ago

3a0165da

Newer Older