Skip to content

fix: Bound reranker input length and attention memory per batch - #836

Draft
buildingjoshbetter wants to merge 4 commits into
maintenance/remaining-code-only-v2from
maintenance/reranker-token-budget-v1
Draft

buildingjoshbetter wants to merge 4 commits into
maintenance/remaining-code-only-v2from
maintenance/reranker-token-budget-v1

fix: Use the budget order directly when slicing long rerank pairs

f1f7655
Select commit
Loading
Failed to load commit list.