completion endpoint in streaming mode
#15007

Merged

Fix: flush partial stop string when <EOG> is reached in /completion endpoint in streaming mode #15007

ggerganov merged 1 commit into ggml-org:master from matteoserva:fix_stop_bug

matteoserva requested a review from

ngxson 97 days ago

github-actions added examples

github-actions added server

matteoserva changed the title ~~flush partial stop string when <EOG> is reached in /completion endpoint in streaming mode~~ Fix: flush partial stop string when <EOG> is reached in /completion endpoint in streaming mode 96 days ago

matteoserva force pushed from 1c4809a6 to a39eac11 76 days ago

matteoserva force pushed from a39eac11 to d3dd9e0e 60 days ago

matteoserva force pushed from d3dd9e0e to c16ac974 17 days ago

matteoserva requested a review from

ggerganov 17 days ago

ggerganov approved these changes on 2025-10-20

send partial stop string when <EOG> is reached

821251ee

matteoserva force pushed from c16ac974 to 821251ee 14 days ago

ggerganov merged 8cf6b42d into master 14 days ago

Reviewers

ggerganov

ngxson

Assignees

No one assigned

Labels

examples server

Milestone

No milestone

llama.cpp Fix: flush partial stop string when <EOG> is reached in /completion endpoint in streaming mode #15007 Merged

Fix: flush partial stop string when <EOG> is reached in /completion endpoint in streaming mode #15007

llama.cpp
Fix: flush partial stop string when <EOG> is reached in /completion endpoint in streaming mode
#15007

Merged