strisunshinewentingwang/pytorch

Pull Requests Commits

[dynamo 3.11] enable other torch 3.11 dynamo-related tests (#99180)

williamwen42 committed 1 year ago

aa8dcab1

add Half support for unary ops on CPU (#98493)

jiayisunx committed 1 year ago

d56e1b2f

[BE] Make some simplifications to torch.utils.checkpoint logic (#101193)

soulitzer committed 1 year ago

98f6b815

[fx] change from #users to num_users in graph printout (#101140)

suo committed 1 year ago

e568c5a1

Enhance Composable FSDP cast forward input tests (#100349)

speediedan committed 1 year ago

2c291491

update timm commit (#100931)

ngimel committed 1 year ago

49578913

[dtensor] remove assertions about submesh checks (#101229)

wanchaol committed 1 year ago

3ae612ba

enable dispatch stub for backend PrivateUse1 (#99611)

caizhi-mt committed 1 year ago

bf50180b

update requirements.txt in /docs (#101092)

dingran committed 1 year ago

e98d762f

[dynamo] Activation checkpointing as higher order op (#101028)

anijain2305 committed 1 year ago

de15e740

[vision hash update] update the pinned vision hash (#101230)

pytorchmergebot committed 1 year ago

c5c75aa0

[GHF][BE] Add `__repr__` to `FlakyRule` (#101234)

malfet committed 1 year ago

ce76670c

[GHF][BE] Memoize `read_flaky_rule (#101239)

malfet committed 1 year ago

f0cc535c

Improve error messages in `THPVariable_set_grad` (#100683)

kiersten-stokes committed 1 year ago

47ec9cc2

Fix typos in error message (#101231)

kit1980 committed 1 year ago

02f15262

use const_data_ptr in get_device_pointers (#100997)

dagitses committed 1 year ago

d9cfa046

Warn on failure to end warmup, add explicit api for start of model invocation (#101129)

eellison committed 1 year ago

b9bfc2b2

Thread generator through trunc_normal_ (#100810)

tfsingh committed 1 year ago

74dc2a53

[inductor] fix incorrect strides in copy() decomp, fix hf_LongFormer + hf_BigBird errors (#100115)

bdhirsh committed 1 year ago

4c8ee583

[aot autograd] fix de-dupping metadata computation bug (#100431)

bdhirsh committed 1 year ago

a6b8e69d

[aot_autograd] proper handling for when outputs are aliased but have identical size/stride/offset metadata (#100430)

bdhirsh committed 1 year ago

5651006b

Towards making torch._inductor.ir typed (#100712)

ezyang committed 1 year ago

2c786961

Add IRNode.realize stub with docs (#100710)

ezyang committed 1 year ago

380054eb

[BE]: enable PLE error codes in ruff and fix bugs (#101079)

Skylion007 committed 1 year ago

738ba13b

[MPS] Fix bernoulli for int types (#100946)

malfet committed 1 year ago

b7bf953b

[dtensor] use stack to manage mesh resources (#101202)

wanchaol committed 1 year ago

599ae95d

[IValue] Better handle sparseTensors in extractStorages (#100783)

rohan-varma committed 1 year ago

6d6abba0

[BE] Simplify tests, elaborate testnames in test_optim.py (#101004)

janeyx99 committed 1 year ago

cb94ea60

[SPMD][BE] Remove the legacy tracing code (#100858)

fegin committed 1 year ago

49c8a0ca

Make interpolate_bilinear deterministic using decomposition (#101115)

ezyang committed 1 year ago

c567748e

Newer Older