ML Container

Is a generic pytorch container for jump starting ML projects.

Setup

Template '.env' file:

CONTAINER_USER="your_name"
DEV_CONTAINER_OWNER="your_name"
PROJECT_DIR="ml_container"
JUPYTER_NOTEBOOK_PORT=5990
DEV_CONTAINER_SSH_PORT=5991

Verifying the Setup

Run following line in the container (enter the container with docker exec -it pytorch-dev-dani /bin/zsh)

uv run main.py

Should give following output, if you don't have a GPU it will look slightly different:

Hello from ml-container! You will now observe random torch output!
torch.randn output: tensor([[ 0.8608,  1.6308, -2.1592],
[ 0.5785,  0.7127,  1.3057],
[-1.0466, -0.4929,  0.8938],
[ 0.2899,  0.8590, -1.5846],
[-0.9523,  1.0531,  0.0618]])
GPU availability = True
GPU name = NVIDIA GeForce RTX 3090