julienp79/ occitan-gemma-4-e2b-it-lora-sfttrainer:latest

4 yesterday

Gemma 4 2B fine-tuned on Occitan via standard LoRA (r=16) with SFTTrainer. Q2_K, Q4_K_M, Q5_K_M, Q8_0, f16 quants available.

... /
params
fa5e003d8101 · 126B
{
"num_ctx": 2048,
"stop": [
"<bos>",
"<|turn>",
"<turn|>",
"<|turn>user"
],
"temperature": 0.7
}