3 Commits

Author	SHA1	Message	Date
Yushun Xiang	242f00a876	add functions to merge images with validation. (#27) ... * add functions to copy images with validation and check the video frame number. * update README to clarify image file handling and maintain metadata structure * Update utils/dataset_merging/merge_lerobot_dataset.py [feat] accelerate the episode file finding with the ` episode_file_mapping`. * Refactor image and video handling in dataset merging tool with enhanced validation * [fix] If `images` directory exists, then use`early_validation` * fix: Import `encode_video_frames` for video encoding in early validation * fix: Add image copying option and update README for PNG support * fix: Remove unused cv2 import from dataset merging utility --------- Co-authored-by: Qizhi Chen <tavish9.chen@gmail.com> Co-authored-by: zhipeng tang <2444198418@qq.com>	2025-05-21 14:43:56 +08:00
Chopin	56850ca928	Add convert stats in parallel using multi-process and decord video backend (#18) ... * add convert_stats.py, convert_dataset_v20_to_v21.py and video_utils.py * fix bugs when num_frame=1 * update readme.md --------- Co-authored-by: Tavish <tavish9.chen@gmail.com>	2025-04-20 00:08:13 +08:00
zhipeng tang	1cdc2a2f40	Create utils ... Delete utils Create merge_lerobot_dataset.py Create README.md Update README.md Update README.md update utils' structure Co-authored-by: Tavish9.chen@gmail.com	2025-04-15 21:10:32 +08:00