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.

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

Make FinetuningStreamingDataset parameters more flexible #1580

Merged
merged 38 commits into from
Oct 14, 2024
Merged

Make FinetuningStreamingDataset parameters more flexible #1580

merged 38 commits into from
Oct 14, 2024

Conversation

XiaohanZhangCMU
Copy link
Member

@XiaohanZhangCMU XiaohanZhangCMU commented Oct 9, 2024 •
edited
Loading

Refactor build_text_dataloader and build_finetuning_dataloader to make them more flexible in terms of accepting extra keys used by StreamingDataset. The rest behavior should stay the same. For example, if the supplied extra keys to the dataloader are not defined in StreamingDataset, a normal TypeError is still expected to be thrown.

@XiaohanZhangCMU XiaohanZhangCMU changed the title Add Streaming-private arguments to LLM-foundry Add more allowed dataset keys to LLM-foundry Oct 9, 2024
@XiaohanZhangCMU XiaohanZhangCMU changed the title Add more allowed dataset keys to LLM-foundry Make FinetuningStreamingDataset parameters more flexible Oct 10, 2024
@XiaohanZhangCMU XiaohanZhangCMU marked this pull request as ready for review October 11, 2024 06:21
@XiaohanZhangCMU XiaohanZhangCMU requested a review from a team as a code owner October 11, 2024 06:21
milocress
milocress reviewed Oct 11, 2024
View reviewed changes
Copy link
Contributor

@milocress milocress left a comment

Choose a reason for hiding this comment

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

Deferring to @dakinggg , my intuition is there is some way of programmatically inspecting the streaming arguments and that this is preferable to hardcoding streams, packing_ratio, etc. as acceptable streaming arguments.

llmfoundry/data/finetuning/dataloader.py Outdated Show resolved Hide resolved
tests/data/test_dataloader.py Outdated Show resolved Hide resolved
dakinggg
dakinggg approved these changes Oct 13, 2024
View reviewed changes
Copy link
Collaborator

@dakinggg dakinggg left a comment

Choose a reason for hiding this comment

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

lgtm

@XiaohanZhangCMU XiaohanZhangCMU merged commit 4a47b5d into main Oct 14, 2024
9 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@milocress milocress milocress left review comments

@dakinggg dakinggg dakinggg approved these changes

@irenedea irenedea Awaiting requested review from irenedea

Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
3 participants
@XiaohanZhangCMU @milocress @dakinggg