Fine-Tuning Workflow¶

from unsloth import FastLanguageModel

model, tokenizer = FastLanguageModel.from_pretrained(
    model_name="unsloth/Qwen2.5-1.5B-Instruct",
    max_seq_length=2048,
    load_in_4bit=True,  # QLoRA
)

model = FastLanguageModel.get_peft_model(
    model,
    r=16,  # LoRA rank
    target_modules=["q_proj", "k_proj", "v_proj", "o_proj"],
    lora_alpha=16,
    lora_dropout=0,
)

from transformers import Trainer

trainer = Trainer(
    model=model,
    tokenizer=tokenizer,
    train_dataset=dataset,
    args=training_args,
)

trainer.train()

model.save_pretrained_gguf(
    "my_finetuned_model",
    tokenizer,
    quantization_method="q4_k_m",  # Recommended for T4
)