Recognize Anything model(RAM) is an image tagging model, which can recognize any common category with high accuracy.
Grounded Segment Anything Model(GSAM) combines Grounding DINO with Segment Anything
Scene Explain explains the image. Different models with different options are available.
- ignore this if you're using Dockerfile
- Create a .env file similar to .env.example
- Get replicate API key from here
- Get scene explain API key from here
- Fill in some secret key
python run.py
In the root dir(btb_ai/), run
docker build -t btb_ai:version .
docker run -p 8000:5000 btb_ai:version
For current version, check CHANGELOG.md
run this at the root dir
pytest
pending...