Dev/eric/agent #43
Merged
Dev/eric/agent #43
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
…ng, implement parallel loading, and enhance filtering with Executor. Update VLM model to Qwen/Qwen2.5-VL-7B-Instruct across relevant components.
… input/output file handling, and creating a dedicated output directory for analysis results.
…o RoboDM format using Ray. Introduce DROIDProcessor class for managing trajectory processing, and update the main execution flow for improved efficiency. Modify droid_vlm_demo to streamline VLM prompt for success detection.
…omponents, streamline droid_vlm_demo by removing unused methods and comments, and enhance dataset processing with lazy loading. Adjust filtering logic and improve output handling for better performance.
…cy instead of F1 score. Update method names, print statements, and summary output accordingly for clarity and consistency.
added 20 commits
July 10, 2025 06:16
…gging Face cache. Modify VLM prompt in droid_vlm_demo for improved specificity in task description.
…by adding checks for TFDS data availability and ensuring valid trajectory data before file creation. Update camera serial mapping logic and adjust dataset loading to use 'droid_100'. Modify .gitignore to include 'droid_100' directory.
…ant text labels and enhancing the VLM prompt for clarity in calibration analysis.
… pipeline usage examples, including auto-scan and quick mode options for trajectory processing.
…cess DROID trajectory directories. Update README to reflect new usage instructions, including automatic ground truth generation and enhanced VLM processing capabilities.
- Updated `.gitignore` to include `eval_runs/`. - Introduced `make_image_grid` function for creating tiled grid images from a list of RGB images. - Enhanced `process_single_trajectory` to support different methods for passing frames to VLM: either as a stream or as a concatenated grid. - Modified `VLMService` to analyze multiple images together with a single prompt. - Updated command-line arguments to allow configuration of frame sampling and passing method. - Improved documentation and comments for clarity on new functionalities.
…eters - Removed image_key and language_key parameters from VLM processing functions as they are not applicable for DROID directories with MP4 files. - Updated the pipeline to handle multiple trials for VLM evaluations, including saving per-trial metrics and aggregate results. - Simplified the example usage in `simple_vlm_processing.py` to reflect the new input format and removed state visualization functionality. - Enhanced documentation to clarify the new input requirements and processing methods.
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.
Summary 📝
Write an overview about it.
Details
Describe more what you did on changes.
Bugfixes 🐛 (delete if dind't have any)
Checks