-
Notifications
You must be signed in to change notification settings - Fork 1.6k
Open
Description
System Info / 系統信息
python3.11 daiban系统
Who can help? / 谁可以帮助到您?
No response
Information / 问题信息
- The official example scripts / 官方的示例脚本
- My own modified scripts / 我自己修改的脚本和任务
Reproduction / 复现过程
`tools = [
{
"name": "base_info",
"description": """查询指定影城的基础信息,包括影院开业时间、影厅数量、今年票房、排名、排座、历年票房表现等等一切影城维度的数据""",
"parameters": {
"type": "object",
"properties": {
"business_date_start": {
"description": "要查询的开始时间,格式为'yyyymmdd'的八位数字,非必传,可以为空",
"type": 'int'
},
"business_date_end": {
"description": "要查询的结束时间,格式为'yyyymmdd'的八位数字,非必传,可以为空",
"type": 'int'
},
"cinema_code": {
"description": "筛选的影城编码,非必传,可以为空",
"type": 'str'
},
"cinema_name": {
"description": "筛选的影城名称,非必传,可以为空",
'type': 'int'
},
"hasService": {
"description": "是否包含服务费,包含为1,不包含为0,非必传,默认为1",
"type": 'str'
},
"periodTime": {
"description": "时段 1:全天 2:黄金场,非必传,默认为1",
"type": 'str'
},
},
"required": ['business_date_start', 'business_date_end', 'cinema_code', 'cinema_name', 'hasService',
'periodTime']
}
},
]
system_info = {
"role": "system",
"content": "Answer the following questions as best as you can. You have access to the following tools:",
"tools": tools
}
messages = [system_info]
messages.append({"role": "system", "content": "不要假设或猜测传入函数的参数值。如果用户的描述不明确,请要求用户提供必要信息"})
messages.append({"role": "system", "content": "不要假设或猜测函数的返回值。如果无法理解返回值的含义,请将返回值完整展示给用户,并告诉用户无法理解,如果函数没有返回任何值,请告诉用户函数调用没有结果"})
messages.append({"role": "system", "content": "如果用户提供的信息不足以调用函数,请告诉用户需要的信息"})
messages.append({"role": "system", "content": "如果用户提供的信息没有涉及非必传参数,请使用默认值调用函数,并告诉用户已使用默认值"})
from transformers import AutoTokenizer, AutoModel
tokenizer = AutoTokenizer.from_pretrained(model_path, trust_remote_code=True)
model = AutoModel.from_pretrained(model_path, trust_remote_code=True, )
model = model.eval()
history = messages
# response, history = model.chat(tokenizer, '查询2024年影城编码为31081501的影院全国排名', history=messages)
for response, history in model.stream_chat (tokenizer, '查询2024年影城编码为31081501的影院全国排名', history=history):
# response, history = model.chat(tokenizer, response, history=history)
print(response)`
Expected behavior / 期待表现
期待修复
Metadata
Metadata
Assignees
Labels
No labels