Spaces:
Sleeping
Sleeping
File size: 6,349 Bytes
e9ce3e8 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 |
import base64
import numpy as np
from typing import Dict
import random
import asyncio
import logging
import os, json
from typing import Any
from aiohttp import ClientSession
from tqdm.asyncio import tqdm_asyncio
import random
from time import sleep
import aiolimiter
import openai
from openai import AsyncOpenAI, OpenAIError
from anthropic import AsyncAnthropic
async def _throttled_openai_chat_completion_acreate(
client: AsyncOpenAI,
model: str,
messages,
temperature: float,
max_tokens: int,
top_p: float,
limiter: aiolimiter.AsyncLimiter,
json_format: bool = False,
n: int = 1,
):
async with limiter:
for _ in range(10):
try:
if json_format:
return await client.chat.completions.create(
model=model,
messages=messages,
temperature=temperature,
max_tokens=max_tokens,
top_p=top_p,
n=n,
response_format={"type": "json_object"},
)
else:
return await client.chat.completions.create(
model=model,
messages=messages,
temperature=temperature,
max_tokens=max_tokens,
top_p=top_p,
n=n,
)
except openai.RateLimitError as e:
print("Rate limit exceeded, retrying...")
sleep(random.randint(10, 20)) # 增加重试等待时间
except openai.BadRequestError as e:
print(e)
return None
except OpenAIError as e:
print(e)
sleep(random.randint(5, 10))
return None
async def generate_from_openai_chat_completion(
client,
messages,
engine_name: str,
temperature: float = 1.0,
max_tokens: int = 512,
top_p: float = 1.0,
requests_per_minute: int = 100,
json_format: bool = False,
n: int = 1,
):
# https://chat.openai.com/share/09154613-5f66-4c74-828b-7bd9384c2168
delay = 60.0 / requests_per_minute
limiter = aiolimiter.AsyncLimiter(1, delay)
async_responses = [
_throttled_openai_chat_completion_acreate(
client,
model=engine_name,
messages=message,
temperature=temperature,
max_tokens=max_tokens,
top_p=top_p,
limiter=limiter,
json_format=json_format,
n=n,
)
for message in messages
]
responses = await tqdm_asyncio.gather(*async_responses)
empty_dict = {
"question": "",
"options": {
"A": "",
"B": "",
"C": "",
"D": "",
},
"distractors": {
"E": "",
"F": "",
"G": "",
},
"correct_answer": ""
}
empty_str = ""
outputs = []
for response in responses:
if n == 1:
if json_format:
if response and response.choices[0] and response.choices[0].message and response.choices[0].message.content:
outputs.append(json.loads(response.choices[0].message.content))
else:
outputs.append(empty_dict)
else:
if response and response.choices[0] and response.choices[0].message and response.choices[0].message.content:
outputs.append(response.choices[0].message.content)
else:
outputs.append(empty_str)
else:
if json_format:
outputs.append([
json.loads(response.choices[i].message.content) if response and response.choices[i].message.content else empty_dict
for i in range(n)
])
else:
outputs.append([
response.choices[i].message.content if response and response.choices[i].message.content else empty_str
for i in range(n)
])
return outputs
async def _throttled_claude_chat_completion_acreate(
client: AsyncAnthropic,
model: str,
messages,
temperature: float,
max_tokens: int,
top_p: float,
limiter: aiolimiter.AsyncLimiter,
):
async with limiter:
try:
return await client.messages.create(
model=model,
messages=messages,
temperature=temperature,
max_tokens=max_tokens,
top_p=top_p,
)
except:
return None
async def generate_from_claude_chat_completion(
client,
messages,
engine_name: str,
temperature: float = 1.0,
max_tokens: int = 512,
top_p: float = 1.0,
requests_per_minute: int = 100,
n: int = 1,
):
# https://chat.openai.com/share/09154613-5f66-4c74-828b-7bd9384c2168
delay = 60.0 / requests_per_minute
limiter = aiolimiter.AsyncLimiter(1, delay)
n_messages = []
for message in messages:
for _ in range(n):
n_messages.append(message)
async_responses = [
_throttled_claude_chat_completion_acreate(
client,
model=engine_name,
messages=message,
temperature=temperature,
max_tokens=max_tokens,
top_p=top_p,
limiter=limiter,
)
for message in n_messages
]
responses = await tqdm_asyncio.gather(*async_responses)
outputs = []
if n == 1:
for response in responses:
if response and response.content and response.content[0] and response.content[0].text:
outputs.append(response.content[0].text)
else:
outputs.append("")
else:
idx = 0
for response in responses:
if idx % n == 0:
outputs.append([])
idx += 1
if response and response.content and response.content[0] and response.content[0].text:
outputs[-1].append(response.content[0].text)
else:
outputs[-1].append("")
return outputs |