Ubuntu Container for Data Science

This repo includes scripts for building an Ubuntu Docker container for data science. It includes commonly used data science and machine learning tools and packages e.g., R, Conda, Python, Spark, Ollama, etc. The details can be found in install_bash_packages.sh.

You can commit the container as an image, so you can pull the image to have a new Ubuntu environment when you want to test something out.

Support both Arm64 and X64 CPU architecture.

Getting Started

# make sure Docker is on
git clone --depth 1 https://github.com/maxleungtszchun/Ubuntu-for-Data-Science.git
cd ./Ubuntu-for-Data-Science
chmod +x ./ubuntu4ds.sh
sudo ./ubuntu4ds.sh

After inputting the above code to your terminal, you only need to choose your username and password and wait for the scripts to complete.

Container to Image

sudo docker commit ubuntu4ds ubuntu4ds_image

Example

sudo docker run -it -p 8080:8080 -p 7860:7860 -w /home/<your-username> --name ubuntu4ds ubuntu4ds_image su <your-username>

# use Ollama to chat with LLM (Llama3.1 8b)
llama

# use fabric to chat with LLM (Llama3.1 8b)
echo 'what is llama3.1' | fabric -sp ai

cat > ~/ModelFile <<-'EOF'
	FROM llama3.1:latest
	PARAMETER num_ctx 15000
	PARAMETER temperature 0
EOF
ollama create llama3.1_ctx_15000 -f ~/ModelFile

sudo nala update && sudo nala install -y poppler-utils
curl -fsSL https://upload.wikimedia.org/wikipedia/commons/1/1a/HKFactSheet_BasicLaw_122014.pdf -o ~/example.pdf
pdftotext ~/example.pdf - | fabric --model llama3.1_ctx_15000:latest -sp extract_wisdom

# you can also visit http://localhost:8080 to use Llama3.1 with Open Webui
# you can even use Stable Diffusion model in Open Webui by:
# setting -> admin settings -> images -> image generation engine = automatic1111 -> Base URL = http://localhost:7860 -> turn on Image Generation (Experimental) -> save

Pull Image directly from Docker Hub

You can pull the image directly from Docker Hub without building.

# for Arm64 CPU
sudo docker run -it -p 8080:8080 -p 7860:7860 -w /home/user --name ubuntu4ds "maxleung414/ubuntu4ds:latest" su user

# for X64 CPU
sudo docker run -it -p 8080:8080 -p 7860:7860 -w /home/user --name ubuntu4ds "ml414/ubuntu4ds:latest" su user