Add the modified openpose and add confidence level filtering.
FangSen9000
adbfc5d
Add optimized ControlNeXt-SVD-v2 implementation with direct pose input support
FangSen9000+1: Claude
6b0f4c6
Major improvements:
- Add run_controlnext_optimized.py with FP16 support, xformers, and GEGLU optimization
- Add direct pose video/images input to bypass DWPose extraction
- Implement smart output naming based on input filename
- Add ComfyUI-ControlNeXt-SVD reference implementation for analysis
- Update README with optimized inference examples
- Add easy_dwpose integration for pose extraction
- Generate example pose videos with text2pose system
Performance optimizations:
- Default FP16 precision for faster inference
- Memory-efficient attention with xformers
- Chunked VAE decoding to reduce memory usage
- GEGLU inplace operations for memory savings
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
Refactor to use individual NPZ files instead of merged database
FangSen9000+1: Claude
ac71c07
Major changes:
- Replace merged database approach with individual NPZ files per video
- Add WLASL_train.json mapping file and 9575 individual NPZ files in pose_dict/
- Create new text2pose_individual.py for working with individual NPZ files
- Add generated_pose_video/ directory for output videos
- Remove old pipeline scripts that worked with merged database
- Update file structure to be more modular and efficient
Benefits:
- More flexible video selection
- Easier to add/remove individual videos
- Better memory efficiency (load only needed videos)
- Cleaner file organization
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
