Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Congratulations on the release!
I'm in the process of seeing how to add DSINE as an alternative surface normal network. In the meantime, I wanted to add Pixi as an alternative installation for this repo as it makes installation + dependency management MUCH easier.
Why?
Pixi combines the conda and pypi environment in a reproducible and easy way, The biggest benefit is that instead of having to go through the long process of Nerfstudio install, having a pixi.toml file takes care of the entire installation with a single command (include cuda dependencies!!). It also includes a .lock file to make sure the environment is reproducible.
How to test
I added a small task that downloads omnidata checkpoints + mushroom sequence + starts training dn-splatter model,
all that's required is having pixi installed (more info here, tldr run
curl -fsSL https://pixi.sh/install.sh | bash
and source your shell)pixi run train-example
To use all commands as if inside a conda environment (that is what pixi uses under the hood) just use
pixi shell
This will install everything and activate the environment
I have personally tested this on two different machines (one with a 1080ti and another one with an A6000) and has worked fine for both!
Let me know if this is helpful or any other questions. Also incase you don't want to add a
pixi.toml
file Pixi recently added support for using pyproject.toml without needing a pixi.toml but I haven't quite gotten that working yet