Replies: 1 comment
-
|
Thanks for the note; made an issue (#1342) to track this. I'm going to close the discussion here, please follow up on the issue. |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
This is an error report and how to evade when following recently updated fine-tuning tutorial
Tutorial page:
https://fair-chem.github.io/core/common_tasks/fine_tuning.html
Always thanks for creating well-behaved MLIP, and diverse datasets for chemistry.
A tutorial for fine-tuning uma model is updated. It helps me a lot to know how-to-play with uma.
There was an error in following tutorial, so I write this page.
Generating toy dataset has no problem, it works fine!
But running a fine-tuning with the prepared dataset, same as ($ fairchem -c uma_sm_finetune_template.yaml)
It gives not a directory error.
As the error looks like to originated from clusterscope, I tried to run clusterscope in cli,
$ cscope info
And almost same error found.
As 'slurm' is not installed in my linux account and 'scontrol' or 's-something' is a word usually used in slurm, It might be related with slurm.
For evading this error, I editted a 171 line on fairchem/src/fairchem/core/_cli.py
Commenting out clusterscope.cluster() and defining cluster to any string solves the error, successfully run fine-tuning.
I think it is not the best method, or I could be wrong at following the fine-tuning tutorial.
But if there is anyone who have trouble in 'scontrol', wish it could help.
Thanks again for passion and contribution in chemistry.
Beta Was this translation helpful? Give feedback.
All reactions