ModelScope爆显存了吗，应该怎么解决？我使用的免费示例，显示GPU内存剩余22G，加载Yi-34B-Chat-4bits模型，kernal总是重启
from modelscope import AutoModelForCausalLM, AutoTokenizer, GenerationConfig

tokenizer = AutoTokenizer.from_pretrained("01ai/Yi-6B-Chat", revision='master', trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("01ai/Yi-6B-Chat", revision='master', device_map="auto", trust_remote_code=True).eval()

Prompt content: "hi"

messages = [
{"role": "user", "content": "hi"}
]
input_ids = tokenizer.apply_chat_template(conversation=messages, tokenize=True, add_generation_prompt=True, return_tensors='pt')
output_ids = model.generate(input_ids.to('cuda'))
response = tokenizer.decode(output_ids[0][input_ids.shape[1]:], skip_special_tokens=True)

Model response: "Hello! How can I assist you today?"

print(response)