strisunshinewentingwang/pytorch

Pull Requests Commits

[profiler] Fix profiling shapes with PT2 + lists of dynamic shapes (#105893)

davidberard98 committed 1 year ago

76fb72e2

[jit] move get_annotations out of infer_concrete_type_builder (#105197)

davidberard98 committed 1 year ago

6938494d

Fix error message about enabling InferenceMode in Python (#105948)

copyrightly committed 1 year ago

22f93852

[FSDP] Add `record_function` for explicit prefetching (#105985)

awgu committed 1 year ago

c099b800

Revert "Simplify handle indexing (#105006)" (#105984)

awgu committed 1 year ago

a9a3c456

fix check issue for replace_params_with_constants (#105909)

XiaobingSuper committed 1 year ago

854fe470

Merge and improve torch optim optimizer type stubs (#102593)

ringohoffman committed 1 year ago

0616952d

[vision hash update] update the pinned vision hash (#105988)

pytorchupdatebot committed 1 year ago

1b9faf22

Bump pygments from 2.12.0 to 2.15.0 in /.github/requirements (#105669)

dependabot[bot] committed 1 year ago

Verified 46b74ab9

[quant][fx] Fix docs for prepare_fx/prepare_qat_fx (#105979)

jerryzh168 committed 1 year ago

d767cff7

[pt2] add meta for `_adaptive_avg_pool3d_backward` (#105816)

nkaretnikov committed 1 year ago

0c65a2d5

Update matmul decomp to match eager (#105850)

lezcano committed 1 year ago

36ae3596

Fix the docs for cosine_similarity (#104772)

lezcano committed 1 year ago

9bde7f4e

Fuse ops in eager cosine_similarity while keeping the stability and the gradients (#104771)

lezcano committed 1 year ago

fff4a9db

test_linalg: triangular_solve - make well_conditioned well conditioned (#105919)

nikitaved committed 1 year ago

a61a0fe4

Add support for tensor.tolist() for static sized int tensors (#105976)

voznesenskym committed 1 year ago

aabdd2b7

fix building errors on FreeBSD (#105897)

cyyever committed 1 year ago

5c5eece6

inductor: fix CSE issue when have symbolic shape input at the freezing path (#105651)

XiaobingSuper committed 1 year ago

afd621dd

inductor: using binary folding path to do conv+bn folding (#105650)

XiaobingSuper committed 1 year ago

9c1802f8

dynamo: support disable_saved_tensors_hooks (#104869)

kshitij12345 committed 1 year ago

920b446d

Delete unused experimental export (#105873)

tugsbayasgalan committed 1 year ago

7b31732a

[Composable] Add more sharding strategies to runtime test (#105205)

rohan-varma committed 1 year ago

03e2ca9d

composable fsdp, checkpoint, + compile test (#105180)

rohan-varma committed 1 year ago

a326f562

[Composable] Use non-reentrant generator, remove reentrant (#105176)

rohan-varma committed 1 year ago

5d70fe01

[dynamo] make ProcessGroupVariable a DistributedVariable (#105593)

wanchaol committed 1 year ago

c76c84bd

[ONNX] Fix the warnings of `aten overload fallback to default` in onnx dispatcher (#105972)

titaiwangms committed 1 year ago

15442915

[profiler] add option for kineto synchronization events in the trace (#105187)

briancoutinho committed 1 year ago

8d9c8897

Don't alter original node's meta in Interpreter (#105880)

SherlockNoMad committed 1 year ago

a770295a

Revert "Disable torchrec/sparse from top-level Dynamo tracing (#105733)"

pytorchmergebot committed 1 year ago

6dd4b99e

Unconditionally record when FakeTensorMode is allocated and report it on inconsistency (#105927)

ezyang committed 1 year ago

884cd53e

Newer Older