Skip to content

Conversation

@ggerganov
Copy link
Member

fix #16876

Update sanity checks in llama-batch to allow repeating positions for mrope embedding inputs.

Copy link
Collaborator

@ngxson ngxson left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks!

JamePeng referenced this pull request in JamePeng/llama-cpp-python Oct 31, 2025
@ggerganov ggerganov merged commit 8da3c0e into master Oct 31, 2025
68 checks passed
@ggerganov ggerganov deleted the gg/batch-fix-pos-check branch October 31, 2025 11:50
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Eval bug: QwenVL models "have inconsistent sequence positions" decode failure when n_img_batches > 1

3 participants