mirror of https://github.com/julyx10/lap.git synced 2026-10-02 01:54:57 +08:00

Files

T

History

julyXX b4229ad9fd fix: release workflow - use native runners, fix matrix format, add Linux deps

- Fix setup job matrix format: use single-line JSON + jq -c to avoid GITHUB_OUTPUT multiline error
- Use macos-15-intel runner for x86_64-apple-darwin (native Intel, no cross-compile)
- Add missing Linux build deps: nasm, clang, pkg-config, libssl-dev, etc.
- Move download scripts to scripts/ folder (download_models.sh + download_models.ps1)
- Remove Windows target (to be added later)
- Change macOS conditions to contains() to match both macos-latest and macos-15-intel
- Replace python with bash for model downloads

2026-02-13 21:30:58 +08:00

download_models.ps1

fix: release workflow - use native runners, fix matrix format, add Linux deps

2026-02-13 21:30:58 +08:00

download_models.sh

fix: release workflow - use native runners, fix matrix format, add Linux deps

2026-02-13 21:30:58 +08:00

README.md

fix: release workflow - use native runners, fix matrix format, add Linux deps

2026-02-13 21:30:58 +08:00

README.md

Scripts

Utility scripts for the Lap project.

download_models

Downloads AI model files (CLIP + InsightFace) required for the app to src-tauri/resources/models/. Already downloaded files are skipped automatically.

Usage

macOS / Linux:

./scripts/download_models.sh

Windows (PowerShell):

.\scripts\download_models.ps1

Models Downloaded

Model	Source	Purpose
`tokenizer.json`	OpenAI CLIP ViT-B/32	Text tokenization
`text_model.onnx`	CLIP ViT-B/32 (quantized)	Text embedding
`vision_model.onnx`	CLIP ViT-B/32 (quantized)	Image embedding
`det_500m.onnx`	InsightFace Buffalo-S	Face detection
`w600k_mbf.onnx`	InsightFace Buffalo-S	Face recognition