oamazonasgabriel/ lfm2.5-8b-a1b:q4_k_m-8gbGPU

1 19 hours ago

A general-purpose 8.3B Mixture-of-Experts model from Liquid AI that activates only ~1.5B parameters per token — delivering strong reasoning, tool calling, and multilingual support while fitting comfortably in 8 GB VRAM.

... /
params
0dcbb3f76ace · 241B
{
"frequency_penalty": 0,
"min_p": 0,
"num_batch": 512,
"num_ctx": 32768,
"num_gpu": 99,
"num_thread": 8,
"presence_penalty": 0,
"repeat_penalty": 1.05,
"stop": [
"im_start",
"im_end",
"tool_call_start",
"tool_call_end"
],
"temperature": 0.2,
"top_k": 80,
"top_p": 0.9
}