- Move all Edge AI Platform code into edge-ai-platform/ subdirectory
- Remove legacy local_service_win/ and relay-server-linux binary
- Keep docs/ and README.md at repo root
- Update docs to latest PRD v3.1 and TDD v2.0
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Merge installer improvements, CI workflow, and full frontend codebase.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
# Conflicts:
# README.md
- Auto-generate random relay token during installation
- Pre-fill relay URL and dashboard URL with EC2 defaults
- Fix hardware detection duplicate device parsing
- Add Dashboard URL field to relay config step
- Launch Server now auto-opens dashboard URL with token
- Add ad-hoc codesign to Makefile installer target
- Remove binary from git tracking
- Add Gitea Actions CI workflow for macOS + Windows builds
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Implemented a new script `video_inference_viewer.py` for processing video files and performing inference using a specified model.
- Added functionality to encode frames in various formats and send them to a specified inference server.
- Included methods for decoding model outputs, applying non-maximum suppression (NMS), and drawing bounding boxes on the video frames.
- Integrated command-line arguments for configuration, including model ID, input dimensions, and output options.
- Added a graphical file dialog for selecting video files.
- Added new images: bike_cars_street_224x224.bmp and one_bike_many_cars_800x800.bmp.
- Updated TEST_PAIRS.md to reflect new source paths for YOLOv5 and FCOS models.
- Introduced VERSION files for KL520 (2.2.0), KL630 (SDK-v2.5.7), and KL720 (2.2.0).
- Added firmware binaries for KL520, KL630, KL720, and KL730.
- Updated third-party Kneron_DFUT resources, including new binaries and translations.
- Added KneronPLUS-1.2.1.zip to third-party resources.
