Skip to content

Conversation

rowellz
Copy link

@rowellz rowellz commented Aug 1, 2024

Made a docker compose environment so users or contributors can more easily get H3DGS running on their own machines. It isn't the most ideal docker configuration, but the pay off is definitely worth it since all you need to run is docker-compose up and the environment will be running in a docker container. I tested this by using my own dataset and shelling into the container with docker exec -it hierarchical_3d_gaussians /bin/bash and successfully ran these commands on my own data:

python ./preprocess/generate_colmap.py --project_dir data/my_data

python ./preprocess/generate_chunks.py --project_dir data/my_data

python ./preprocess/generate_depth.py --project_dir data/my_data

python ./scripts/full_train.py --project_dir data/my_data

@rowellz
Copy link
Author

rowellz commented Aug 14, 2024

Hi @White-Mask-230 thanks for reviewing/approving this, any idea what else needs to happen before merging this? I would like to add glomap to the Dockerfile to speed up the colmap process & to maybe work on a backwards compatible approach with exhaustive feature matching, but that is probably out of scope for this.

@White-Mask-230
Copy link

I don't think it's necessary to add anything more

Fabulani referenced this pull request in Fabulani/hierarchical-3d-gaussians Feb 17, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants