strisunshinewentingwang/pytorch

Pull Requests Commits

[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

Implement coalesced all_gather_into_tensor (#101157)

kwen2501 committed 1 year ago

daed3bf8

[BE] Testing docs: clarify test instantiation function usage (#100905)

jbschlosser committed 1 year ago

e47cdd0c

Remove obsolete upsample_bilinear2d lowerings (#101111)

ezyang committed 1 year ago

ae233286

sparse compressed validation: allow empty-batched inputs (#101180)

nikitaved committed 1 year ago

346e1f51

Fix incorrect sparse_dim in COO.zero_() and in binary operations with zero-sized COO operands (#98292)

pearu committed 1 year ago

65b15be0

Update torchbench pin (#101071)

ezyang committed 1 year ago

41a4e220

Add global boolean for controlling whether to record concrete shapes or not (#101043)

davidberard98 committed 1 year ago

f7571507

Add is_xpu to torch type (#101072)

guangyey committed 1 year ago

14964b3a

[pt2] Skip if curr_size is None (#101170)

xw285cornell committed 1 year ago

1e6002be

Newer Older