strisunshinewentingwang/pytorch

Pull Requests Commits

Print AOT Autograd graph name when accuracy failed (#99366)

ezyang committed 1 year ago

c17ff0ed

[vmap] Fix searchsorted batch rule for self_logical_rank == 0 (#99526)

qqaatw committed 1 year ago

47215534

[BE] `stdint.h`->`cstdint` (#99570)

malfet committed 1 year ago

2ad02d00

Revert "[vmap] Fix searchsorted batch rule for self_logical_rank == 0 (#99526)"

pytorchmergebot committed 1 year ago

35ad5122

inductor(CPU): add ISA check before do cpu fx packed weight (#99502)

XiaobingSuper committed 1 year ago

ccd5ad81

Revert "Temporarily move ROCm to unstable (#99579)"

pytorchmergebot committed 1 year ago

4d890688

Revert "Fix trailing spaces lint (#99581)"

pytorchmergebot committed 1 year ago

21e88a54

Revert "Allow in graph einops operators (#99478)"

pytorchmergebot committed 1 year ago

96a262d6

[functorch] Prevent using for-loop for out-of-place index_fill batch rule (#99229)

qqaatw committed 1 year ago

edd2507c

[FSDP]Make param_groups optional for FSDP optim state dict (#99117)

wz337 committed 1 year ago

a2a41442

[inductor] a script to benchmark the perf impact from tensor layout (#99583)

shunting314 committed 1 year ago

68bc0fc0

Enable torch.jit.load for custom device (#99535)

heidongxianhua committed 1 year ago

da322ea8

[vmap] Fix searchsorted batch rule for self_logical_rank == 0 (#99526)

qqaatw committed 1 year ago

6580b160

[vmap] Add max_pool3d batch rule (#99522)

qqaatw committed 1 year ago

c0674c43

[vamp] Add max_pool1d batch_rule (#99517)

qqaatw committed 1 year ago

d31a00e7

fix onednn ConvTranspose2d channels last issue when ic=1 (#99539)

XiaobingSuper committed 1 year ago

233cc34d

inductor: support sqrt for dynamic shape (#99514)

XiaobingSuper committed 1 year ago

3af467ef

inductor: add input type check for fuse_attention (#99296)

XiaobingSuper committed 1 year ago

27a43c02

Allow in graph einops operators (#99478)

wconstab committed 1 year ago

309b7edf

[ONNX] Update install_onnx.sh: onnx-script -> onnxscript (#99572)

justinchuby committed 1 year ago

95ca8e58

[Dynamo] Implementing generic context manager by inlining __enter__ and __exit__ (#98725)

yanboliang committed 1 year ago

78907098

Fix trailing spaces lint (#99581)

kit1980 committed 1 year ago

fbdb86c1

Temporarily move ROCm to unstable (#99579)

kit1980 committed 1 year ago

d06624d3

Add an expect test for test_save_graph_repro (#99538)

ezyang committed 1 year ago

805a6dc8

[quant][pt2e][refactor] Move the annotation for observer sharing ops into separate util (#99384)

jerryzh168 committed 1 year ago

36acad58

increment generation in run only context (#99099)

eellison committed 1 year ago

1b9edb68

Adding vmap support for special bessel functions (#99543)

SiddharthIVEX committed 1 year ago

c660db20

Remove a couple of additional places where we would construct tensors - aliases of params, inputs (#98950)

eellison committed 1 year ago

19788002

Add TorchXLA option to benchmark runner (#99505)

jansel committed 1 year ago

3233450d

Make HAS_CPU boolean lazy, speed up import time (#99537)

ezyang committed 1 year ago

6026caed

Newer Older