Morph MORPH
SIGN IN SIGN UP
dhruvhead / pytorch UNCLAIMED
0 0 121 Python

COMMITS

main
August 22, 2026
D
[sdpa] Fix bottom-right efficient attention when Q exceeds K (#193632)
drisspg committed
a698136
August 24, 2026
L
Fix symbolic FakeTensor support for shift out overloads (#194375)
Lucas Bordeaux committed
445878d
I
[MPS] Raise NotImplementedError for pooling ops that require 64-bit indexing (#194460)
Isalia20 committed
84c32f8
P
Revert "Prevent saving mutated grad inputs in AOT min-cut (#185876)"
PyTorch MergeBot committed
20ef2cc
Y
Fix CUDA illegal memory access in embedding_bag per_sample_weights backward (#193408)
Yunqian Fan committed
550fa84
P
Revert "[BC breaking] Remove assert_allclose (#164560)"
PyTorch MergeBot committed
9bb4e0c
P
[Test] Refactor `test_dynamic_shapes.py` (#194081)
Piotr Bielak committed
0fee942
P
[Test][XPU] Enable XPU for `test_autocast.py` (#193189)
Piotr Bielak committed
8bd1b7f
P
[Test][XPU] Unskip test cases in `test_modules.py` (#193384)
Piotr Bielak committed
179d721
P
[1/N][Test] Add hw_classification to `TestSparseCSRSampler` class (#193170)
Piotr Bielak committed
6f6183e
P
Skip fake autocast tests for `linalg.pinv` and `pinverse` on XPU (#178277)
Piotr Bielak committed
f772fda
Y
[BC breaking] Remove assert_allclose (#164560)
Yuanyuan Chen committed
eac5354
N
[XPU][Aten] Fix memory format detection for conv backward overrideable (#194328)
NaOHCC committed
c240daa
P
[Testcase Refactoring] Enable PrivateUse1 tests and add hw-classification in test_microbatch (#190422)
PlatformLC committed
f426618
F
[CUDA] Remove unused multinomial probability copy (#194391)
fxlin0261 committed
5f7293f
P
[Testcase Refactoring] Use is_backend_available for MultiProcContinuousTest backend check (#192186)
PlatformLC committed
373d690
T
[Testcase Refactoring] Add hw_classification to test_store (#192036)
twthighd committed
8b1669d
P
[torchtitan hash update] update the pinned torchtitan hash (#194477)
PyTorch UpdateBot committed
7f098f9
P
[torchcomms hash update] update the pinned torchcomms hash (#194551)
PyTorch UpdateBot committed
5a9e9bc
P
[vllm hash update] update the pinned vllm hash (#194552)
PyTorch UpdateBot committed
d1cbd4b
D
[Test][Refactor] Decouple test_dataloader according to the accelerator relevance. (#189250)
daisyden committed
8a531fc
C
Disable USE_FLASH_ATTENTION when no CUDA arch supports it (#194502)
cyy committed
68d20d4
August 21, 2026
F
Fix typos and grammatical errors in comments and docstrings (#194322)
Frederik Gossen committed
6a0eb75
August 23, 2026
J
Fix Dynamo tracing of compile support checks (#185528)
Jason Ansel committed
e26e727
J
Fix AttrProxy module reuse in Dynamo (#185300)
Jason Ansel committed
3b0b5dc
F
Fix typos and grammar in comments and docstrings (#194492)
Frederik Gossen committed
3574fb4
F
Fix grammar and typos in code comments and docstrings (#194493)
Frederik Gossen committed
f335d6c
F
Fix grammar and typos in code comments and docstrings (#194494)
Frederik Gossen committed
24abbdc
J
Fix SymBool equality handling in symbolic shapes (#186461)
Jason Ansel committed
b7930e4
J
Enable torch.compile for RNN modules by default (#185902)
Jason Ansel committed
a2ed3ae