[feat] support video evaluation for qwen2-vl and add mix-evals-video2text #275
Conversation
This file contains 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
…anage samples saving process
Eliminated the commented-out import statement for WandbLogger to tidy up the code and enhance readability. This helps maintain focus on active components and prevents confusion over unused code. A cleaner structure contributes to better maintainability in the long run. No functional changes were made, just a step towards a more streamlined codebase.
- Integrated vision processing for videos and images, improving context handling within the model. - Added error logging for missing utility dependencies to inform users about installation requirements. - Updated YAML configuration to standardize prompt handling for various video tasks. - Bumped version number to indicate ongoing development status. These changes streamline how visuals are managed in the model, contributing to better assistant responses in tasks involving media.
kcz358
approved these changes
Sep 23, 2024
|
Now Qwen2-VL is changed to use original Here's the difference:
|
- Added automatic naming for W&B runs if not specified, improving organization. - Updated video frame rate from 1.0 to 0.5 for better performance and resource management during visual content processing. - Streamlined W&B logging by removing redundant code, ensuring cleaner execution flow. These changes optimize logging efficiency and enhance the overall user experience.
- Updated chat template logic for better formatting in responses, ensuring consistent handling of user and assistant roles. - Reduced maximum new tokens in multiple evaluation files to ensure more concise outputs and improve efficiency. - Enhanced clarity in few-shot tasks by explicitly labeling question and answer roles in generated text. - Simplified logging of contextual and target information during evaluation, ensuring better tracking of results. These adjustments improve the overall output quality and streamline the evaluation processes.
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.
These changes streamline how visuals are managed in the model, contributing to better assistant responses in tasks involving media.