Skip to content

Support chat templatization and tool calling in SFT datasets #864

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.

Sign up for GitHub

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

Jump to bottom
Open
wants to merge 6 commits into
base: main
Choose a base branch
from
Open

Support chat templatization and tool calling in SFT datasets #864

wants to merge 6 commits into from

Conversation

ananthsub
Copy link
Contributor

fix #838
fix #837

  • Adds HuggingFace chat template support through use_hf_tokenizer_chat_template
  • Tool calling & function schemas
  • OpenAI messages format conversion
  • Packed sequence support with chat templates
  • Pass dataset_kwargs through packing pipeline
  • update old copy/pasted docstrings & typehints

based heavily on:

@ananthsub ananthsub self-assigned this Oct 3, 2025
@copy-pr-bot copy-pr-bot
Copy link

copy-pr-bot bot commented Oct 3, 2025

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

@ananthsub
Copy link
Contributor Author

/ok to test 777dd78

@ananthsub
Copy link
Contributor Author

/ok to test b908582

@ananthsub
sync 13367
75419e8 
Signed-off-by: Ananth Subramaniam <[email protected]>
@ananthsub
add tool calling and templatization support
c0632d1 
Signed-off-by: Ananth Subramaniam <[email protected]>
@ananthsub
more tests
32c748b 
Signed-off-by: Ananth Subramaniam <[email protected]>
@ananthsub
only include attention mask in batch if needed
076b33c 
Signed-off-by: Ananth Subramaniam <[email protected]>
@ananthsub
safe map
86242b2 
Signed-off-by: Ananth Subramaniam <[email protected]>
@ananthsub
fix test
84eacbc 
Signed-off-by: Ananth Subramaniam <[email protected]>
@ananthsub ananthsub force-pushed the tool-calling-dataset branch from b908582 to 84eacbc Compare October 3, 2025 17:55
@ananthsub
Copy link
Contributor Author

/ok to test 84eacbc

@ananthsub ananthsub marked this pull request as ready for review October 3, 2025 20:51
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@dimapihtar dimapihtar Awaiting requested review from dimapihtar

@cuichenx cuichenx Awaiting requested review from cuichenx

At least 1 approving review is required to merge this pull request.

Assignees

@ananthsub ananthsub

Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.

Chat dataset support missing from MegatronBridge Tool calling support missing from MegatronBridge Datasets
1 participant
@ananthsub