New LFM2.5 8b A1b model!!
Posted by Glittering_Focus1538@reddit | LocalLLaMA | View on Reddit | 8 comments
Performance is on par with Nemotron 3 Nano, at an even higher speed!
I will be adding support to SmallCode for this model as it uses non-standard tool calls.
AdInternational5848@reddit
Yea what am I supposed to do with this?
AiGenom@reddit
Which tasks its can make?
Truth-Does-Not-Exist@reddit
it's not worth testing guy's
Specific_Dimension51@reddit
Why ?
Glittering_Focus1538@reddit (OP)
It's pretty dumb, only good for basic tool calls and small tasks.
Covert-Agenda@reddit
Just downloaded it, rapid on a m5 max mbp
Educational_Mud4588@reddit
Lots of posts for this model. Completely enjoying it on an intel cpu system without vram and getting 20 tg/s
Glittering_Focus1538@reddit (OP)
getting 230tok/s on my rx9070 It's dumb but fast, its okay for simple tasks.