RosettaCommons / RosettaCommons/RFdiffusion
installation on read only directory :: problems
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 3.1k
- Forks
- 644
- PR merge metrics
- No merged PRs in 30d
Description
Hello,
I was asked to install RFdiffusion on our central cluster.
RFdiffsuion was installed in a python-3.10 virtual env
git clone https://github.com/RosettaCommons/RFdiffusion.git
pip install necessary requireements (see https://github.com/RosettaCommons/RFdiffusion/issues/95)
then pip install RFdiffusion
installation is done on a read only nfs drive.
so running run_inference.py from the compute node leads to some permission errors as by default it tries eg to cached IGSO3 to PREFIX/lib/python<PYTHON_VERSION>/site-packages/schedules which is RO
same may apply to models directory that are by default searched in PREFIX/lib/python<PYTHON_VERSION>/site-packages/models, ie impossible for user to put the wheight models in where exepected.
I understand that actual code is OK for a user that runs run_inference.py from his own space with RW rights ;-)
may I suggest that by default 'sensible' directories, eg config, models, schedules cache dir will be located and searched in $HOME directory.
regards
Contributor guide
No contributing guide indexed for this repository
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start with run_inference.py and inspect how the config, models, and schedules paths are selected and cached. Reproduce the permission errors from a read-only Python environment, then verify that a user-writable location such as $HOME can be used without changing the installation. Confirm the inference workflow works on the compute node with the required weight models.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- devops, machine-learning
- Issue type
- Feature
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 42/100