Open johanneskruse opened 4 months ago
Hi Johannes
We used Kaggle Docker to ensure consistency in our environments:
Please follow these steps to replicate our environments:
Pull the Kaggle Docker image from the URL provided below. The Docker image version we tested was v126, but most recent versions should work without major issues.
https://console.cloud.google.com/artifacts/docker/kaggle-gpu-images/us/gcr.io/python
docker pull gcr.io/kaggle-gpu-images/python:v126
Launch a Docker container using this image.
docker run --gpus all -it --rm \
-p 8888:8888 \
-v /path/to/your/local/directory:/home/
gcr.io/kaggle-gpu-images/python:v126
Please install the polars package if your docker image does not have polars. Since polars implemented destructive changes from version 1.0.0, you should install a version from the 0.x series, such as 0.18.4.
pip install polars==0.18.4
Access Jupyter Lab and execute the ipynb files in sequence.
jupyter lab --ip=0.0.0.0 --port=8888 --allow-root
As for docker execution environment, we use Linux/Ubuntu. GPU is required for computing embeddings.
Hi all,
Could you specify the dependencies required for the project to run. Make a requirements.txt, pyproject.toml, or poetry.toml. As well as specify the OS you used.