siyandaV-0/multi-modal-docker-image

Docker Image for jupyter-lab environment for multi-modal models (stable diffusion, etc.)

Multi-Modal Models Docker Image:

---

• This image builds a multi-modal environment for use in jupyter-lab
  
• The multimodality in question is mostly diffusion architectures.
  
• The image builds opencv from source code to have cuda support.
  
• You can then run the cv algorithms using the GPU for faster processing.
  
• This will serve you well if you can deploy the container to a remote server.
  

Working Directory:

Our working Directory is structured as follows:

.
└── computer-vision-docker-image
    ├── dependecies
    │   ├── mm-requirements.txt
    │   ├── ubuntu-deps.sh
    │   └── opencv.sh
    ├── test-code
    │   ├── controlnet.ipynb
    │   └── stable-diffusion.ipynb    
    ├── Dockerfile
    ├── docker-compose.yml
    └── ReadME.md

To build the image run the docker-build bash script

 docker-compose build

If you need to make updates to the content of the container, e.g. installing new packages with pip- you can specify these in the cv-requirements.txt file in the dependencies folder.

---

To launch the image run the docker-run bash script

 docker-compose up

The docker run bash scripts allow for ease of launching into the container's jupyter-lab.
This docker run script uses docker-compose commands to achieve this.