qwen2-7b-instruct-traininfer-instruct0625/merge_weight.py

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

model_id = "Qwen/Qwen2-7B-Instruct"
tokenizer = AutoTokenizer.from_pretrained(model_id, trust_remote_code=True)
output_path = "Qwen2-7B-Instruct-Merged"

model = AutoModelForCausalLM.from_pretrained(
    model_id,
    device_map="auto",
    torch_dtype=torch.float16,
    trust_remote_code=True,
).eval()

model = PeftModel.from_pretrained(model, model_id="./ckpt")
merged_model = model.merge_and_unload()
merged_model.save_pretrained(
    output_path, max_shard_size="2048MB", safe_serialization=True
)