strisunshinewentingwang/pytorch

Pull Requests Commits

Make sure we dealloc on recording, not just replay (#97440)

eellison committed 1 year ago

f1b901b0

[CUDA12] set_device change (#94864)

Aidyn-A committed 1 year ago

c18be2b2

Fix GridSample Activation Quantization (#98278)

leimao committed 1 year ago

7b088890

Add test for pickle_module (#98373)

malfet committed 1 year ago

3da7e832

add new() method identifier to _StorageBase (#98201)

XDaoHong committed 1 year ago

ea00f850

Only print guard code when printing guards (#98347)

ezyang committed 1 year ago

2a32bc50

Add itemsize and nbytes properties to Tensor (#98322)

bjhargrave committed 1 year ago

555ab310

fix _slice_meta's shape calculation (#98326)

ar0ck committed 1 year ago

14ccad73

Fix complex variable notation for division operator to be consistent. (#98057)

redwrasse committed 1 year ago

b4420f0f

Uniformly use elem when checking ListType (#97873)

dilililiwhy committed 1 year ago

526b564f

[CI] Mark sebotnet33ts_256 as nondeterministic (#98356)

desertfire committed 1 year ago

c4de7fde

[inductor] Fix slow tests not being run in CI (#97841)

peterbell10 committed 1 year ago

a05d787e

Revert "remove typed StorageImpl::data() and StorageImpl::unsafe_data() (#98219)"

pytorchmergebot committed 1 year ago

45edc58e

Move android-emulator-build-test to periodic (#98370)

kit1980 committed 1 year ago

752e43c3

Inductor cpp wrapper: support dynamic shapes (#97965)

chunyuan-w committed 1 year ago

2987bc07

Fix typos under caffe2/operators directory (#98235)

kiszk committed 1 year ago

601e7dc0

Account for forwards which whose corresponding backwards are not invoked (#98112)

eellison committed 1 year ago

feb9ec42

Fix saving and loading pickle files on Big Endian systems (#95881)

AlekseiNikiforovIBM committed 1 year ago

ae0d06b4

Revert "[spmd expansion] support torch.ops.aten.sym_numel (#98229)" (#98382)

weiwangmeta committed 1 year ago

1e3abda3

Fix incorrectly getting the name of OrderedDict's index in dynamo (#96940)

knwng committed 1 year ago

e943b120

Do not track parameters, do not generate guards (#98350)

ezyang committed 1 year ago

30d47e45

remove typed StorageImpl::data() and StorageImpl::unsafe_data() (#98219)

mikey dagitses committed 1 year ago

144d5268

[inductor] Fix a perf regression caused by https://github.com/pytorch/pytorch/pull/98214 (#98343)

desertfire committed 1 year ago

6887333c

Switch accuracy CI to dynamic batch only (#98307)

ezyang committed 1 year ago

b923f848

Fix typos under torch/distributed directory (#98225)

kiszk committed 1 year ago

6514d71a

remove typed StorageImpl::unsafe_data() (#98218)

mikey dagitses committed 1 year ago

3af02283

Increment pending forwards after invocation (#98101)

eellison committed 1 year ago

a3365e1d

[SyncBatchNorm] Support running with low precision parameters (#98332)

awgu committed 1 year ago

3686416a

Extend TensorImpl with BackendMeta (#97429)

ppiskorski committed 1 year ago

2d9b2bcf

Revert "[pt2][inductor] add `triton.__verison__` as cache key, update cache layout (#98010)"

pytorchmergebot committed 1 year ago

dd503376

Newer Older