fix(generation): handle CUDA multinomial limit in beam search sampling#45369
Closed
sharziki wants to merge 1 commit intohuggingface:mainfrom
Closed
fix(generation): handle CUDA multinomial limit in beam search sampling#45369sharziki wants to merge 1 commit intohuggingface:mainfrom
sharziki wants to merge 1 commit intohuggingface:mainfrom