Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Print conversation/sample if an error occurrs processing dataset #593

Closed
5 tasks done
teknium1 opened this issue Sep 17, 2023 · 1 comment
Closed
5 tasks done

Print conversation/sample if an error occurrs processing dataset #593

teknium1 opened this issue Sep 17, 2023 · 1 comment
Labels
enhancement New feature or request

Comments

@teknium1
Copy link
Contributor

⚠️ Please check that this feature request hasn't been suggested before.

  • I searched previous Ideas in Discussions didn't find any similar feature requests.
  • I searched previous Issues didn't find any similar feature requests.

🔖 Feature description

When processing the dataset to tokenize, and an error occurrs for whatever reason, print to terminal the last example loaded to process, to help diagnose dataset issues.

✔️ Solution

print(example)

❓ Alternatives

No response

📝 Additional Context

No response

Acknowledgements

  • My issue title is concise, descriptive, and in title casing.
  • I have searched the existing issues to make sure this feature has not been requested yet.
  • I have provided enough information for the maintainers to understand and evaluate this request.
@teknium1 teknium1 added the enhancement New feature or request label Sep 17, 2023
@NanoCode012
Copy link
Collaborator

Since we're calling map from datasets library and we use multiple sub-classes of PromptTokenizingStrategy which implement their own tokenize prompt, I'm not sure how we could add a general try catch around everything and still get the exact prompt that caused the issue.

https://github.com/OpenAccess-AI-Collective/axolotl/blob/c1382e79b6eb599f7fd210fd1a621b97bfed0d51/src/axolotl/datasets.py#L41-L45

However, I think this is a useful feature to have!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
enhancement New feature or request
Projects
None yet
Development

No branches or pull requests

3 participants