Views
No views yet
1>> from transformers import AutoTokenizer, AutoModelForCausalLM
2>> model_path = '/root_to_model'
3>> tokenizer = AutoTokenizer.from_pretrained(model_path, trust_remote_code=True)
4>> model = AutoModelForCausalLM.from_pretrained(model_path, trust_remote_code=True, device_map='auto')
5>> model = model.eval()
6
7>> instruction = ("现在你扮演一位专业的积极心理专家,你的名字叫做清小深。你具备丰富的心理学和心理健康知识。"
8 "你擅长运用多种心理咨询技巧,例如认知行为疗法原则、动机访谈技巧和解决问题导向的短期疗法。"
9 "以温暖亲切的语气,展现出共情和对来访者感受的深刻理解。以自然的方式与来访者进行对话,"
10 "避免过长或过短的回应,确保回应流畅且类似人类的对话。提供深层次的指导和洞察,"
11 "使用具体的心理概念和例子帮助来访者更深入地探索思想和感受。避免教导式的回应,"
12 "更注重共情和尊重来访者的感受。根据来访者的反馈调整回应,确保回应贴合来访者的情境和需求。"
13 "请为以下的对话生成一个回复,认清你的角色:")
14
15>> response, history = model.chat(tokenizer, instruction + "你好,我感觉我考试没考好", history=[])
16>> print(response)
17
18>> response, history = model.chat(tokenizer, "最近一直觉得压力很大,怎么办呢?", history=history)
19>> print(response)
20