I've been tracking and working with Open Llama for months. I picked it as my foundational model for other work for a few reasons:
Permissive license without Meta's draconian acceptable use policy.
More flexible model sizes (3B).
Good baseline performance.
I'm getting a bit concerned that this project may be abandoned, though. It's been over a month since the last model release. I'm willing to help pitch in and support additional training for this model. The alternative is that I start up my own foundational model using a similar mix of data. I'd rather support this effort, if possible. Please let me know how I can help!
I've been tracking and working with Open Llama for months. I picked it as my foundational model for other work for a few reasons:
I'm getting a bit concerned that this project may be abandoned, though. It's been over a month since the last model release. I'm willing to help pitch in and support additional training for this model. The alternative is that I start up my own foundational model using a similar mix of data. I'd rather support this effort, if possible. Please let me know how I can help!