strisunshinewentingwang/pytorch

Pull Requests Commits

[7/n][FSDP] make use_dtensor=True work with offload_to_cpu=True for load_state_dict (#105378)

wz337 committed 1 year ago

c54f6302

Create tensor from Numpy in current device. (#105546)

ysiraichi committed 1 year ago

5ce5372d

[BE] Enable ruff's UP rules and autoformat test/ (#105434)

justinchuby committed 1 year ago

73e14553

fix typo (#105507)

metamomo committed 1 year ago

7b562385

Update pybind11 submodule to 2.11.0 (#105245)

Skylion007 committed 1 year ago

801fb93b

[EZ][BE] Fix the massively annoying strict-weak-ordering issue. (#105189)

debowin committed 1 year ago

70b5264e

[ONNX] Add missing spaces between sentences in warning text (#105527)

tomaarsen committed 1 year ago

4448c78a

switching NNC as default for TorchScript support (#105185)

jjsjann123 committed 1 year ago

218b5477

[composable API] Fix the replicate_device_id test case to avoid copy replicated models. (#105503)

fegin committed 1 year ago

a10f93f6

[dynamo] add initial dynamo support for DTensor (#103146)

wanchaol committed 1 year ago

f139aab2

Switch UFMT to denylist rather than allowlist (#105536)

ezyang committed 1 year ago

a788365d

[BE] Enable ruff's UP rules and autoformat distributed/ (#105433)

justinchuby committed 1 year ago

232b96b6

[MPS][BE] Use `Tensor::copy_` (#105475)

malfet committed 1 year ago

2125794c

[BE] Enable ruff's UP rules and autoformat dynamo / functorch and refs (#105432)

justinchuby committed 1 year ago

8a688277

Upgrade nightly wheels to rocm5.6 (#105076)

BLOrange-AMD committed 1 year ago

88f11977

[BE] Enable ruff's UP rules and autoformat inductor/ (#105431)

justinchuby committed 1 year ago

cb7a30f6

[BE] Enable ruff's UP rules and autoformat ao/ (#105430)

justinchuby committed 1 year ago

c0d8a4af

Improve validator module behavior if Z3 is not installed. (#105168)

ysiraichi committed 1 year ago

0b6de0eb

[dynamo][torch_np] support linalg, random and fft module (#105320)

kshitij12345 committed 1 year ago

e137ac6c

inductor: promote half/bfloat16 constant to float for cpu vectorization path (#105440)

XiaobingSuper committed 1 year ago

18bcf62b

Fix for "AttributeError when attempting to remove inductor buffers twice" (#104901)

mixarcid committed 1 year ago

7ddb66e3

[inductor] Suport OMP on MacOS (#105136)

nkflash committed 1 year ago

167eab1c

Use shareable calculate-docker-image GHA (#105372)

huydhn committed 1 year ago

0e85c224

[quant][pt2e][be] Rename prepare_pt2e_quantizer to prepare_pt2e (#105484)

jerryzh168 committed 1 year ago

554052f3

[BE] Enable ruff's UP rules and autoformat benchmarks/ (#105429)

justinchuby committed 1 year ago

5ef023b0

[pytorch] Change autograd fallback mode to Nothing (#105505)

zou3519 committed 1 year ago

9c225c9b

Produce more logs from TCPStore init (#105350)

XilunWu committed 1 year ago

d2fa3f60

Fix mps unary op issue on non densely stored tensors (#105512)

Mr4k committed 1 year ago

d2c24eca

Upgrade triton pin (#105463)

ezyang committed 1 year ago

133a5e9a

Fix a docstring of resolve_neg (#104151)

kato8966 committed 1 year ago

64c39ece

Newer Older