Regarding my paper submission before i submit

[Richard-Selvaraj-99]

Hello i have a project that i can submit as paper as i have done a experimental analysis all my code and project is available via github and hugging space its a experimental analysis on llama 3.1 how ever the reproducable code and the experiemtation code is under mit licencse but not the llama 3.1itself as it does not meet osi guideleines would it be valid to submit this paper ?

[warrickball]

JOSS is intended as a venue for code that is designed to be re-used by other researchers in your domain for their own analyses, not just to reproduce an existing analysis. Note also that we have requirements about the overall size and scope of the codebase (it must represent "substantial scholarly effort") and that our docs also say

While useful for many areas of research, pre-trained machine learning models and notebooks are not in-scope for JOSS.

Using external libraries that are not open source isn't fundamentally a barrier to submission but your code itself must be licensed under an OSI-approved licence.

It's difficult to decide on a specific case without seeing the submission but to me this does not sound like something in scope for JOSS.

[Richard-Selvaraj-99]

My code is under joss but what i want to know is thw work is based on llama model which is not open source

…

On Thu, 9 Oct 2025, 8:21 pm Warrick Ball, ***@***.***> wrote: *warrickball* left a comment (openjournals/joss#1468) <#1468 (comment)> JOSS is intended as a venue for code that is designed to be re-used by other researchers in your domain for their own analyses, not just to reproduce an existing analysis. Note also that we have requirements <https://joss.readthedocs.io/en/latest/submitting.html#substantial-scholarly-effort> about the overall size and scope of the codebase (it must represent "substantial scholarly effort") and that our docs also say <https://joss.readthedocs.io/en/latest/submitting.html#what-we-mean-by-research-software> While useful for many areas of research, pre-trained machine learning models and notebooks are not in-scope for JOSS. Using external libraries that are not open source isn't fundamentally a barrier to submission but your code itself must be licensed under an OSI-approved licence. It's difficult to decide on a specific case without seeing the submission but to me this does not sound like something in scope for JOSS. — Reply to this email directly, view it on GitHub <#1468 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/A5V3JU5OVK5P76BHNJNFTAD3W2YTNAVCNFSM6AAAAACIXS46SSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTGOBXGIYDMNRSGM> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[warrickball]

It's difficult to evaluate without a submission. There is precedent in JOSS for publishing submissions where the submitted code is open source but requires the installation of separate publicly-available but not open source software.

Again, though, pre-trained (including fine-tuned) AI/ML models are not in-scope for JOSS, whether or not they (or any thin wrappers) are open source.

JOSS is also not a venue for publishing results from experiments or analyses, even if they are experiments on software.