How to use alexiaassis/Modelo-Treinado-Gemma3 with PEFT:
from peft import PeftModel from transformers import AutoModelForCausalLM base_model = AutoModelForCausalLM.from_pretrained("unsloth/gemma-3-12b-it-unsloth-bnb-4bit") model = PeftModel.from_pretrained(base_model, "alexiaassis/Modelo-Treinado-Gemma3")
ecab80c 63bb3aa ecab80c
1
2
3
4
5
6
7
8
9
{ "epoch": 3.0, "num_input_tokens_seen": 6120032, "total_flos": 4.115769119160883e+17, "train_loss": 0.1491297289889346, "train_runtime": 4379.932, "train_samples_per_second": 2.545, "train_steps_per_second": 0.318 }
