You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Commit 21d9c64
Browse filesBrowse the repository at this point in the historyBrowse files
[ET-VK] Deduplicate transition clone nodes in TagMemoryMetaPass
Pull Request resolved: pytorch#18771
When the same tensor is consumed by multiple ops that need a different
storage representation, the pass previously inserted a separate clone
transition for each consumer. Now it caches transition clones keyed by
(source_node, target_storage_type, target_layout) and reuses existing
clones when the same transition is needed again.
For Qwen3 0.6B (8da4w fp16), the embedding output (BUFFER due to
vocab_size exceeding texture limits) feeds both rms_norm and add which
need TEXTURE. Previously 2 clones were inserted; now 1 clone is shared.
Authored by Claude.
ghstack-source-id: 364280900
@exported-using-ghexport
Differential Revision: [D100004700](https://our.internmc.facebook.com/intern/diff/D100004700/)
0 commit comments