bachelor-oppgave-nina

Maintainers:

• Francesco Frassinelli

Original authors:

• Lars Blütecher Holter
• Benjamin Letnes Bjerken
• Lillian Alice Wangerud
• Daniel Hao Huynh

Requirements

• pixi
• pre-commit

Installation

Download model weights:

pixi run fetch-models

Install pre-commit hooks:

pre-commit install
pre-commit install --hook-type commit-msg

See the commitlint config: https://github.com/conventional-changelog/commitlint/tree/master/%40commitlint/config-conventional

Run Locally

pixi run start

Testing

Run unit tests:

pixi run pytest tests

Run VDI performance tests (requires GPU and test video data):

pixi run test

Run using Docker

docker compose --profile prod build
docker compose --profile prod run --rm app

Directory Structure

• data/input/ - Place input video files (.mp4) here for processing
• data/output/ - Processed video files and detection results are saved here
• data/models/ - YOLO model weights (.pt files)

The performance test automatically picks the first .mp4 file from data/input/ and saves results to data/output/.

Notes

Model weights are automatically downloaded by the fetch-models task, or you can manually download them from GitHub release v0.1.0.