Skip to content

server : pre-calculate EOG logit biases#14721

Merged
ggerganov merged 1 commit intomasterfrom
gg/server-eos-pre-calc
Jul 16, 2025
Merged

server : pre-calculate EOG logit biases#14721
ggerganov merged 1 commit intomasterfrom
gg/server-eos-pre-calc

Conversation

@ggerganov
Copy link
Copy Markdown
Member

cont #14710

Avoid iterating the vocabulary on each request when "ignore_eos" parameter is set. To do this, pre-calculate the EOG tokens in advance.

@ggerganov ggerganov merged commit 6ffd4e9 into master Jul 16, 2025
51 of 56 checks passed
blime4 referenced this pull request in blime4/llama.cpp Feb 5, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants