strisunshinewentingwang/pytorch

Pull Requests Commits

Expand sparse.softmax zero nnz tests to cover cases of previously reported FPE. (#95646)

amjames committed 1 year ago

2bcc0e9e

Use FindCUDAToolkit to find cuda dependencies (#82695)

peterbell10 committed 1 year ago

c5f60925

sparse compressed tensor validation without syncs for low-(batch)dim tensors. (#94048)

nikitaved committed 1 year ago

7901f2d1

[MPS] Fix views with 3 or more sliced dimensions (#95762)

DenisVieriu97 committed 1 year ago

e5a959a2

[CI] Reduce the frequency of running inductor-perf-test-nightly (#95778)

desertfire committed 1 year ago

7d097e36

[CI] Change the way tests are triggered with dynamo and inductor (#94539)

desertfire committed 1 year ago

9835c93a

[inductor] correctly infer dtype of `full` (#95593)

nkaretnikov committed 1 year ago

e3892fd1

[Inductor] Fix the logical_and/logical_or vectorization issue (#95609)

EikanWang committed 1 year ago

9da903f1

[Inductor] Vectorize channels-last adaptive_avg_pool2d (#95608)

EikanWang committed 1 year ago

c1f5e50f

[Inductor] Fix the issue that at::vec does not support indexing (#95459)

EikanWang committed 1 year ago

074ae720

[dtensor] add submesh example to checkpoint_example (#95655)

wanchaol committed 1 year ago

7a772bff

Rearrange some transformer tests (#95745)

drisspg committed 1 year ago

3fa93962

Dynamic dim guards (#95584)

voznesenskym committed 1 year ago

1e2e1495

Don't generate guards that refer to unbacked SymInts (#95732)

ezyang committed 1 year ago

e628a3e7

allow privateuse1 key to be used with legacy constructor (#95748)

bdhirsh committed 1 year ago

9b86b532

raw_values is dead (#95703)

ezyang committed 1 year ago

93f1aa55

Avoid recursion in graph traverse (#95723)

wschin committed 1 year ago

9227fd74

[CI] Compile on M1 natively (#95719)

malfet committed 1 year ago

e970dd9d

[CI] Force clear triton cache between running each test (#95729)

desertfire committed 1 year ago

e79b2b77

[vision hash update] update the pinned vision hash (#95665)

pytorchmergebot committed 1 year ago

d3d75a5c

[inductor] fix type promotion for comparison operations (#95736)

ngimel committed 1 year ago

21b1134b

Small bugfix in nested matmul bmm path head_dim acquisition (#95744)

mikaylagawarecki committed 1 year ago

6930f30c

Fix kernel name bug (#95739)

mlazos committed 1 year ago

e50ff3fc

[Inductor Perf Test Workflow] Remove pull request trigger and rely on ciflow/ label only (#95755)

weiwangmeta committed 1 year ago

65f49ab6

feat(dockerfile): shrink layers & build cleaner (#95375)

georgettica committed 1 year ago

1c526664

Correct OneCycleLR doc example code to explicitly call optimizer.step() (#95730)

kiersten-stokes committed 1 year ago

60a1d295

[MPS] Add support for masked_scatter (#95743)

DenisVieriu97 committed 1 year ago

ed1957dc

[BE][DDPOptimizer] De-dup `p` and `param` (#95654)

awgu committed 1 year ago

d9cd9a13

Initial minifier smoke test + runbook (#95670)

mlazos committed 1 year ago

94bec94f

Remove dead ZeroGuard (#95701)

ezyang committed 1 year ago

7ea3aab4

Newer Older