strisunshinewentingwang/pytorch

Pull Requests Commits

[JIT] Document runtime flags

davidberard98 committed 2 years ago

5316f967

Disable ROCM builds/tests (#80862)

malfet committed 2 years ago

f7678055

Fix rrelu on CUDA (#80434)

lezcano committed 2 years ago

c42c5943

Add support for multiple inputs to out_wrapper and strict dtype checking (#80601)

lezcano committed 2 years ago

eb0889cf

Revert "Add DLPack support for XPU backend by mapping to kDLOneAPI in DLPack. (#78154)"

pytorchmergebot committed 2 years ago

877180e1

[torchdynamo hash update] update the pinned torchdynamo hash (#80852)

pytorchmergebot committed 2 years ago

c78bdd7b

[Convolution] Fix derivatives of convolution overrideable backward (#80840)

Qiza-lyhm committed 2 years ago

fce5560e

In order to make pytorch headers consumable from cpp20 code bases, … (#79985)

Lukas N Wirz committed 2 years ago

5af48581

Revert "Revert "Add support for directly passing symint to empty"" (#79954)

Krovatkin committed 2 years ago

0a5123a7

Add DLPack support for XPU backend by mapping to kDLOneAPI in DLPack. (#78154)

chengjunlu committed 2 years ago

3a6c1dc7

Make `kl_div` a composite function. (#80334)

lezcano committed 2 years ago

828c787e

Add -Werror=non-virtual-dtor (#80584)

huydhn committed 2 years ago

76700358

Improve readability of cuda_lazy_init (#80788)

cbalioglu committed 2 years ago

081b56fd

Exclude Fake dispatch key during tensor construction (#80782)

cbalioglu committed 2 years ago

c54aabf3

[MPS] Add test consistency from OpInfo based tests from PR 78504 (#79532)

kulinseth committed 2 years ago

76cff182

Update docstring for scale_factor in torch.nn.functional.interpolate. (#80807)

dogeplusplus committed 2 years ago

b4ed13ea

[torchdynamo hash update] update the pinned torchdynamo hash (#80825)

pytorchmergebot committed 2 years ago

482e53c6

MPS: Fix handling of 1D tensors in linear backward (#80759)

razarmehr committed 2 years ago

0e3953fc

[torchdynamo hash update] update the pinned torchdynamo hash (#80801)

pytorchmergebot committed 2 years ago

caa6ef15

[ONNX] Convert aten numpy_T to ONNX transpose (#79269)

siboehm committed 2 years ago

b00e4e80

[torchdynamo hash update] update the pinned torchdynamo hash (#80790)

pytorchmergebot committed 2 years ago

88807bcf

Revert "Revert [ONNX] Enable data propagation from ONNX to simplify current shape inference" (#80730)

jcwchen committed 2 years ago

667fdd72

Do not use `thrust::lower_bound` on device (#80746)

malfet committed 2 years ago

8f9e3a20

[jiterator] De-template launch_jitted_reduce_kernel (#80138)

peterbell10 committed 2 years ago

e1cfac42

[jiterator] Reduce templating in jitted_gpu_kernel_impl (#80103)

peterbell10 committed 2 years ago

df665b1a

Expose linalg.solve_ex (#80073)

lezcano committed 2 years ago

19f3d4d7

Improve heuristics for linalg_lu_solve when B is a matrix (#79838)

lezcano committed 2 years ago

6fb6dc42

Make slogdet, linalg.sloget and logdet support metatensors (#79742)

lezcano committed 2 years ago

37a58196

Add scatter support for view operations (#79939)

kulinseth committed 2 years ago

b744e1c8

Transform constraints to z3 constraints which is the final step (#80110)

migeed-z committed 2 years ago

31952b56

Older