strisunshinewentingwang/pytorch

Pull Requests Commits

Expose API to specify custom context manager for checkpoint (#96783)

soulitzer committed 1 year ago

f3db2a63

Add exceptionhandler to more distributed_c10d APIs (#96770)

H-Huang committed 1 year ago

ac7329b3

[CUDA] Use accumulate type to improve accuracy of grid_sample on half precision inputs [v2] (#96586)

xwang233 committed 1 year ago

1716709d

Output peak memory stats from dynamo torchbench perf CI (#95666)

wconstab committed 1 year ago

54cd4a67

Use .to instead of contiguous to generate channels last tensor (#96791)

eellison committed 1 year ago

44586312

[functorch] batch rule : few decomposition ops (#96744)

kshitij12345 committed 1 year ago

24c49dbf

add to functorch codeowner (#96746)

kshitij12345 committed 1 year ago

9b1b3fdd

[doc] fix `torch.cuda.mem_get_info` doc (#96621)

stas00 committed 1 year ago

11e708dd

Temporarily disable test_ops_jit on Windows (#96859)

huydhn committed 1 year ago

6339ee5d

add 2d tests to ci (#96711)

wanchaol committed 1 year ago

aa09f889

Fixes a layer_norm_nested backwards edge case. (#96788)

drisspg committed 1 year ago

5612aa6a

Fix type hint for `torch.Tensor.grad_fn` (#96804)

XuehaiPan committed 1 year ago

80e8e41c

Fix build, shadowed variable (#96778)

eellison committed 1 year ago

a7d2e451

[aot autograd] avoid cloning some inputs unnecessarily when they dont require grad (#96342)

bdhirsh committed 1 year ago

e9d9151e

[aot autograd] refactor to make functionalization self-contained (#96341)

bdhirsh committed 1 year ago

3cd9c7a1

[Optim in backward] register_hook=False API (#95096)

rohan-varma committed 1 year ago

8e628726

aot autograd: consolidate metadata (#96340)

bdhirsh committed 1 year ago

8ce8d49c

aot_autograd: dont requires_grad on tangents (#96339)

bdhirsh committed 1 year ago

070cefae

aot autograd refactor: make all synthetic base logic layered in a single location (#96235)

bdhirsh committed 1 year ago

a2694699

[aot_autograd] only performance functionalization analysis pass once (#95992)

bdhirsh committed 1 year ago

7a076b7b

Revert "[memory profiling] add a facility to gather combined C++/Python/TorchScript stack traces. (#95541)"

pytorchmergebot committed 1 year ago

e1ea584b

[reland][CI] switch torchbench to a pinned version (#96782)

desertfire committed 1 year ago

33c7be36

COO intersection: allow external hash + hash reuse in sparse_mask (#94596)

nikitaved committed 1 year ago

3fd24fb6

[MPS] Fix the MacOS 13.3 selector check. (#96733)

kulinseth committed 1 year ago

93f79969

Bump black version to 23.1.0 (#96578)

BowenBao committed 1 year ago

60a68477

[BE] Enforce sign-compare (#96723)

malfet committed 1 year ago

a229e785

Fix int() casting in torch.nn.RNN to have correctly traced JIT and ONNX graph. (#92970)

borisfom committed 1 year ago

96c745df

Revert "Update xnnpack to the latest commit (#95884)"

pytorchmergebot committed 1 year ago

d3a38bdd

Rework torch.compile docs (#96706)

msaroufim committed 1 year ago

6110effa

Revert "Fix periodic job by excluding check_graph_breaks (#96780)"

pytorchmergebot committed 1 year ago

27952336

Newer Older