Skip to content

Conversation

@kallewoof
Copy link
Contributor

@kallewoof kallewoof commented Aug 14, 2025

This checks whether -np is insufficient vs context, and provides clues as to how much is needed for each.

The current error message is incorrectly blaming insufficient context e.g. for TruthfulQA which requires -np=16 or so due to num_answers being near that value.

This checks whether -np is insufficient vs context, and provides clues as to how much is needed for each.

int num_answers = cur_task.seq_tokens.size();
if (s0 + num_answers > max_seq) {
max_seq_exceeded = s0 + num_answers;
Copy link
Member

Choose a reason for hiding this comment

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

We can log the error here and straight up return. This way no need to declare max_seq_exceeded var

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Done.

Copy link
Contributor Author

@kallewoof kallewoof Aug 14, 2025

Choose a reason for hiding this comment

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

I guess I should check s0 == 0 instead of i0 == i1. Edit: switched to doing that.

@ggerganov ggerganov merged commit 3ea913f into ggml-org:master Aug 14, 2025
46 of 47 checks passed
@kallewoof kallewoof deleted the 202508-better-errors-pplcmd branch August 14, 2025 06:17
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants