Spaces:

---

CodeSoft

/

SLM-Arena

like 13

Running on Zero

I think he only does 100M+ models so its fair, param-wise.

Ah, okay — that makes sense

There isn't a set parameter range, and I’ll add models of any size as long as they’re under around 250M parameters. I’ll add those soon. Thanks for the request!

Thank you very much, I really appreciate it!

There isn't a set parameter range, and I’ll add models of any size as long as they’re under around 250M parameters. I’ll add those soon. Thanks for the request!

Oh, sorry for my ignorance, but wouldn't it be unfair?

There isn't a set parameter range, and I’ll add models of any size as long as they’re under around 250M parameters. I’ll add those soon. Thanks for the request!

Oh, sorry for my ignorance, but wouldn't it be unfair?

The goal is generally to compare AI models regardless of size. The biggest inspiration for this project, arena.ai, doesn't limit models to a specific price/size. At some point I plan to add charts of score vs parameters (or something like that.)

There isn't a set parameter range, and I’ll add models of any size as long as they’re under around 250M parameters. I’ll add those soon. Thanks for the request!

Oh, sorry for my ignorance, but wouldn't it be unfair?

The goal is generally to compare AI models regardless of size. The biggest inspiration for this project, arena.ai, doesn't limit models to a specific price/size. At some point I plan to add charts of score vs parameters (or something like that.)

Yeah, that makes sense, but more often than not, the bigger models will have better generations.

I mean, I'm not against the idea of doing models across parameter sizes; I'm just pointing out it isn't entirely fair.

Anyway, this project is awesome; and I think this could go far.

I just added OpenCerebral/Boris-1.3-125M, OpenCerebral/Boris-1.3-75M, OpenCerebral/Boris-1.3-125M-Instruct, and OpenCerebral/Boris-1.3-75M-Instruct. Thank you for the suggestion!

Thank you so much!