transformers
[Paged-Attention] Handle continuous batching for repetition penalty
#39457
Merged

[Paged-Attention] Handle continuous batching for repetition penalty #39457

ArthurZucker merged 9 commits into main from penalty-cb
kashif
kashif Handle continuous batching for repetition penalty
635bc875
kashif kashif requested a review from ArthurZucker ArthurZucker 273 days ago
HuggingFaceDocBuilderDev
ArthurZucker
ArthurZucker approved these changes on 2025-07-17
kashif fix last scores and with token mask creation
84cf493f
kashif add test
e34f1f49
kashif Merge branch 'main' into penalty-cb
0cfcee20
kashif kashif requested a review from ArthurZucker ArthurZucker 271 days ago
kashif Merge branch 'main' into penalty-cb
622e68bf
ArthurZucker
ArthurZucker commented on 2025-07-21
ArthurZucker
ArthurZucker approved these changes on 2025-07-22
kashif Update src/transformers/generation/continuous_batching.py
4e9cc2f6
kashif Update src/transformers/generation/logits_process.py
2393d7b1
kashif fix formatting
57aa84b9
kashif
kashif commented on 2025-07-22
kashif remove unneeded cast
0909e421
ArthurZucker ArthurZucker merged 2936902a into main 267 days ago
ArthurZucker ArthurZucker deleted the penalty-cb branch 267 days ago
ArthurZucker

Login to write a write a comment.

Login via GitHub

Reviewers
Assignees
No one assigned
Labels
Milestone