Spaces:
Runtime error
Runtime error
File size: 6,732 Bytes
182d290 691ae91 182d290 691ae91 182d290 691ae91 182d290 691ae91 182d290 691ae91 182d290 691ae91 182d290 691ae91 182d290 691ae91 182d290 691ae91 182d290 691ae91 182d290 544fec9 182d290 691ae91 182d290 691ae91 182d290 691ae91 182d290 691ae91 182d290 691ae91 182d290 544fec9 182d290 691ae91 182d290 691ae91 182d290 691ae91 74cf2ac |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 |
from typing import Optional, List
# from langchain.llms.utils import enforce_stop_tokens
# import torch
import requests
# import logging
# from transformers import AutoTokenizer, AutoModel, AutoConfig
# logging.basicConfig(filename='chat_log.txt', level=logging.INFO)
DEVICE = "cuda"
FORWARD_KEY = 'fk198719-Pmvv22OqZiovaxRq6YxCzkTcd6UVVX5O0'
# def torch_gc():
# if torch.cuda.is_available():
# with torch.cuda.device(DEVICE):
# torch.cuda.empty_cache()
# torch.cuda.ipc_collect()
class ChatGLM:
max_length: int = 10000
temperature: float = 0
top_p = 0.9
tokenizer: object = None
model: object = None
history_len: int = 10
history = []
URL = 'http://183.131.3.48:9200'
HEADERS = {'Content-Type': 'application/json'}
@property
def _llm_type(self) -> str:
return "ChatGLM"
def __call__(self,
prompt: str,
history: Optional[List[list[str]]] = None,
stop: Optional[List[str]] = None) -> str:
# print('\n\n\n\n')
# print('-------------------------------------------------------------------------------------------------------')
# print(' ****** prompt ****** ')
# print(prompt)
if history:
history = [i for i in history if i[0] is not None] # clear out the system message
history = history[-self.history_len:]
params = {'tokenizers': self.tokenizer, 'prompt': prompt, 'history': history, 'top_p': self.top_p,
'max_length': self.max_length, 'temperature': self.temperature}
response = requests.post(self.URL, headers=self.HEADERS, json=params).json()
answer = response['response']
# question = prompt.split('question:\n')[-1]
# self.history = self.history+[[prompt, response]]
# print(" ****** GLM_answer ****** ")
# print(answer)
# print('-------------------------------------------------------------------------------------------------------')
# print('\n\n\n\n')
return answer
class LocalChatGLM:
max_length: int = 10000
temperature: float = 0
top_p = 0.9
tokenizer: object = None
model: object = None
history_len: int = 10
history = []
@property
def _llm_type(self) -> str:
return "ChatGLM"
def __call__(self,
prompt: str,
history: List[List[str]] = [],
stop: Optional[List[str]] = None) -> str:
# print('\n\n\n\n')
# print('-------------------------------------------------------------------------------------------------------')
# print('**************** prompt ****************：')
# print(prompt)
response, _ = self.model.chat(
self.tokenizer,
prompt,
history=history[-self.history_len:] if self.history_len > 0 else [],
max_length=self.max_length,
temperature=self.temperature,
)
# torch_gc()
# if stop is not None:
# response = enforce_stop_tokens(response, stop)
question = prompt.split('question:\n')[-1]
self.history = self.history+[[question, response]]
# print("*********************** answer **************************：")
# print(response)
# print('-------------------------------------------------------------------------------------------------------')
# print('\n\n\n\n')
return response
# @classmethod
# def load_model(cls,
# model_name_or_path: str = "THUDM/chatglm-6b"):
# tokenizer = AutoTokenizer.from_pretrained(
# model_name_or_path,
# trust_remote_code=True
# )
# if torch.cuda.is_available() and DEVICE.lower().startswith("cuda"):
# model = (
# AutoModel.from_pretrained(
# model_name_or_path,
# trust_remote_code=True)
# .half()
# .cuda()
# )
# else:
# model = (
# AutoModel.from_pretrained(
# model_name_or_path,
# trust_remote_code=True)
# .float()
# .to(DEVICE)
# )
# llm = cls()
# llm.tokenizer = tokenizer
# llm.model = model
# return llm
class OpenAI3:
max_length: int = 10000
temperature: float = 0.2
top_p = 0.9
tokenizer: object = None
model: object = None
history_len: int = 10
history = []
HEADERS = {'Content-Type': 'application/json', 'Authorization': 'Bearer fk198719-pHAOCyaUXohoZBl0KfRvYf4AuHhWm8pm'}
URL ='https://openai.api2d.net/v1/chat/completions'
MODEL_NAME = "gpt-3.5-turbo"
@property
def _llm_type(self) -> str:
return "OPENAI3"
def __call__(self,
prompt: str,
history: Optional[List[List[str]]] = None,
stop: Optional[List[str]] = None) -> str:
message = [{"role": "user", "content": prompt}]
params = {"model": self.MODEL_NAME, "messages": message, 'temperature': self.temperature}
response = requests.post(self.URL, headers=self.HEADERS, json=params).json()
answer = response['choices'][0]['message']['content']
# if stop is not None:
# answer = enforce_stop_tokens(answer, stop)
return answer
class OpenAI4:
max_length: int = 10000
temperature: float = 0.2
top_p = 0.9
tokenizer: object = None
model: object = None
history_len: int = 10
history = []
HEADERS = {'Content-Type': 'application/json', 'Authorization': 'Bearer fk198719-pHAOCyaUXohoZBl0KfRvYf4AuHhWm8pm'}
URL ='https://openai.api2d.net/v1/chat/completions'
MODEL_NAME = "gpt-4"
@property
def _llm_type(self) -> str:
return "OPENAI4"
def __call__(self,
prompt: str,
history: Optional[List[List[str]]] = None,
stop: Optional[List[str]] = None) -> str:
message = [{"role": "user", "content": prompt}]
params = {"model": self.MODEL_NAME, "messages": message, 'temperature': self.temperature}
response = requests.post(self.URL, headers=self.HEADERS, json=params).json()
answer = response['choices'][0]['message']['content']
# if stop is not None:
# answer = enforce_stop_tokens(answer, stop)
return answer |