Feature/phi4 fine tune #181
Open
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…ding, audio layer removal, and input filtering
…unit tests for vision detection
…ndling for vision-only processing
…_step for Phi-4 model
… a constant tensor for vision mode
…n load_model function
|
|
|
What is required to progress on this? |
|
I think we have to finish the review and do changes if anything is required but otherwise nothing in my opinion |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
Adding support for fine tuning Phi-4 Multimodal instruct
List any dependencies that are required for this change.
Type of change
Please delete options that are not relevant.
How has this change been tested, please provide a testcase or example of how you tested the change?
Testing in progress
Any specific deployment considerations
QLora is not working as the base model is trained with Lora adapters and are merged in the transformer, we might have to monkey patch it or find other way to do it
For example, documentation changes, usability, usage/costs, secrets, etc.
Docs