- Add VLA episode schema for observation-action trajectories.
- Add JSONL writer and recorder utilities.
- Add mock VLA recording demo without hardware.
- Add LeRobot-compatible conversion validation stub.
- Add VLA dataset format, data recording, safety/privacy, and LeRobot integration docs.
- Add future smoke tests for generated VLA episodes.
- Reorganized MLLCV into a low-latency computer vision tracking and robotics perception prototype.
- Added A8 Mini low-latency predictive tracking workflow documentation.
- Added RTSP latest-frame capture pipeline documentation.
- Added YOLO detector integration notes.
- Added AsymTrack tracking workflow notes.
- Added Kalman prediction and visual servo control documentation.
- Added dry-run gimbal mode and smoke-test commands.
- Added GitHub issue templates and pull request template.
- Added lightweight GitHub Actions validation workflow.
- Added
CONTRIBUTING.md,CODE_OF_CONDUCT.md, andSECURITY.md.
- This is an early-stage prototype and requires hardware-specific tuning.
- Private model weights, RTSP URLs, TensorRT engines, and recordings are not included.
- CI intentionally avoids hardware, GPU-only checks, and private assets.