LoRA Fine-Tuning: Train AI Models with Minimal Resources

LoRA Fine-Tuning: Train AI Models with Minimal Resources

💡 What You Will Learn

LoRA Fine-Tuning: Train AI Models with Minimal Resources

LoRAWhat Is

pip install torch transformers datasets accelerate peft

[
  {
    "instruction": "Translate to English",
    "input": "",
    "output": "The weather is nice today."
  },
  {
    "instruction": "",
    "input": "",
    "output": "……"
  }
]

from transformers import AutoModelForCausalLM, AutoTokenizer
from peft import LoraConfig, get_peft_model
import torch

# 4bit
model = AutoModelForCausalLM.from_pretrained(
    "Qwen/Qwen2.5-7B-Instruct",
    load_in_4bit=True,
    torch_dtype=torch.float16
)
tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen2.5-7B-Instruct")

# ConfigurationLoRA
lora_config = LoraConfig(
    r=8,           # 
    lora_alpha=32, # 
    target_modules=["q_proj", "v_proj"],  # 
    lora_dropout=0.05,
    bias="none"
)

model = get_peft_model(model, lora_config)
print(f": {model.num_parameters(only_trainable=True)/1e6:.2f}M")
# Parameter1400.1%

from transformers import TrainingArguments, Trainer

training_args = TrainingArguments(
    output_dir="./lora-output",
    per_device_train_batch_size=4,
    gradient_accumulation_steps=4,
    num_train_epochs=3,
    learning_rate=2e-4,
    fp16=True,
    save_strategy="epoch"
)

trainer = Trainer(
    model=model,
    args=training_args,
    train_dataset=dataset
)
trainer.train()

from peft import PeftModel

base_model = AutoModelForCausalLM.from_pretrained("Qwen/Qwen2.5-7B-Instruct")
lora_model = PeftModel.from_pretrained(base_model, "./lora-output/checkpoint-xxx")
merged = lora_model.merge_and_unload()
merged.save_pretrained("./final-model")

|:----|:-------:|:--------:| | 1.5B | 12GB | 4GB | | 7B | 24GB | 6GB | | 13B | 48GB | 12GB | | 70B | 320GB | 48GB |