LoRA微调7B大模型全流程：数据准备到推理效果对比

pip install torch==2.3.1 --index-url https://download.pytorch.org/whl/cu121
pip install transformers==4.44.2 peft==0.12.0 bitsandbytes==0.43.3 trl==0.9.6 accelerate==0.33.0 datasets==2.20.0

{"instruction": "劳动合同违约金上限是多少？", "output": "根据《劳动合同法》第二十二条和二十三条，只有两种情形可以约定违约金：一是违反服务期约定，二是违反竞业限制约定。除此之外，用人单位不得与劳动者约定由劳动者承担违约金。服务期违约金不得超过用人单位提供的培训费用，且不得超过服务期尚未履行部分所应分摊的培训费用。"}

from datasets import load_dataset
from transformers import AutoTokenizer

model_name = "Qwen/Qwen2.5-7B-Instruct"
tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)
tokenizer.pad_token = tokenizer.eos_token

def format_example(example):
    messages = [
        {"role": "system", "content": "你是一名专业法律顾问，请根据中国现行法律法规回答问题。"},
        {"role": "user", "content": example["instruction"]},
        {"role": "assistant", "content": example["output"]}
    ]
    text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=False)
    return {"text": text}

dataset = load_dataset("json", data_files={"train": "train.jsonl", "valid": "valid.jsonl"})
dataset = dataset.map(format_example, remove_columns=dataset["train"].column_names)

def tokenize_fn(examples):
    tokens = tokenizer(examples["text"], truncation=True, max_length=1024, padding=False)
    tokens["labels"] = tokens["input_ids"].copy()
    return tokens

tokenized = dataset.map(tokenize_fn, batched=True, remove_columns=["text"])

import torch
from transformers import AutoModelForCausalLM, BitsAndBytesConfig
from peft import LoraConfig, get_peft_model, prepare_model_for_kbit_training

bnb_config = BitsAndBytesConfig(
    load_in_4bit=True,
    bnb_4bit_quant_type="nf4",
    bnb_4bit_use_double_quant=True,
    bnb_4bit_compute_dtype=torch.bfloat16
)

model = AutoModelForCausalLM.from_pretrained(
    model_name,
    quantization_config=bnb_config,
    device_map="auto",
    trust_remote_code=True
)
model = prepare_model_for_kbit_training(model)
model.gradient_checkpointing_enable()

lora_config = LoraConfig(
    r=64,
    lora_alpha=128,
    lora_dropout=0.05,
    bias="none",
    task_type="CAUSAL_LM",
    target_modules=["q_proj","k_proj","v_proj","o_proj","gate_proj","up_proj","down_proj"]
)

model = get_peft_model(model, lora_config)
model.print_trainable_parameters()
# 输出：trainable params: 40,370,176 || all params: 7,655,986,688 || trainable%: 0.5273

from transformers import TrainingArguments
from trl import SFTTrainer

training_args = TrainingArguments(
    output_dir="./qwen2.5-7b-lora-legal",
    per_device_train_batch_size=2,
    per_device_eval_batch_size=2,
    gradient_accumulation_steps=8,
    num_train_epochs=3,
    learning_rate=2e-4,
    lr_scheduler_type="cosine",
    warmup_ratio=0.03,
    logging_steps=10,
    eval_strategy="steps",
    eval_steps=50,
    save_strategy="steps",
    save_steps=100,
    save_total_limit=3,
    bf16=True,
    optim="paged_adamw_8bit",
    gradient_checkpointing=True,
    report_to="none",
    max_grad_norm=0.3,
    group_by_length=True
)

trainer = SFTTrainer(
    model=model,
    args=training_args,
    train_dataset=tokenized["train"],
    eval_dataset=tokenized["valid"],
    tokenizer=tokenizer,
    dataset_text_field="text",
    max_seq_length=1024,
    packing=False
)

trainer.train()
trainer.save_model("./qwen2.5-7b-lora-legal/final")

from peft import PeftModel
from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

base_model = AutoModelForCausalLM.from_pretrained(
    "Qwen/Qwen2.5-7B-Instruct",
    torch_dtype=torch.bfloat16,
    device_map="auto"
)
model = PeftModel.from_pretrained(base_model, "./qwen2.5-7b-lora-legal/final")
model = model.merge_and_unload()  # 合并LoRA权重，推理更快
tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen2.5-7B-Instruct")

def chat(question):
    messages = [
        {"role": "system", "content": "你是一名专业法律顾问。"},
        {"role": "user", "content": question}
    ]
    text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
    inputs = tokenizer(text, return_tensors="pt").to(model.device)
    outputs = model.generate(**inputs, max_new_tokens=256, temperature=0.3, do_sample=True)
    return tokenizer.decode(outputs[0][inputs["input_ids"].shape[1]:], skip_special_tokens=True)

print(chat("竞业限制补偿金标准是多少？"))