Comments (7)
我检查了自己的log 日志,发现请求一直失败:
2023-09-11 14:23:24,354 Request 5-0 failed with Exception None
2023-09-11 14:23:24,355 Starting request #5-0
2023-09-11 14:23:24,355 Request 5-0 failed with Exception None
2023-09-11 14:23:24,357 Starting request #5-0
2023-09-11 14:23:24,358 Request 5-0 failed with Exception None
不知道是什么原因造成的
from llm4rs.
Same question with huazhen02, 似乎是没有连接成功主机,尝试用作者提供的proxy跑一遍
from llm4rs.
大佬使用的proxy似乎不work,尝试用大佬的curl指令测试网络结果如下:
(env_name) [bianz@login001 LLM4RS]$ curl https://api.openai.com/v1/chat/completions \
-H "Content-Type: application/json"
-H "Authorization: Bearer sk-xxxxxxx"
-d '{
"model": "gpt-3.5-turbo",
"messages": [{"role": "user", "content": "Hello!"}]
}'
{
"id": "chatcmpl-7xna0bYORjlclyvn0tzi7JWQC3Dha",
"object": "chat.completion",
"created": 1694486292,
"model": "gpt-3.5-turbo-0613",
"choices": [
{
"index": 0,
"message": {
"role": "assistant",
"content": "Hello! How can I assist you today?"
},
"finish_reason": "stop"
}
],
"usage": {
"prompt_tokens": 9,
"completion_tokens": 9,
"total_tokens": 18
}
}
请问大佬这种结果意味着什么?需要如何处理
from llm4rs.
- 一般跑多久
这个要根据max_requests_per_minute
、max_tokens_per_minute
和你的网络情况而定。在不触及openai发送请求的上限情况下,max_requests_per_minute
和max_tokens_per_minute
这俩参数越大,跑得越快。script/run.py
的例子,网络情况良好的情况下,半小时跑完绰绰有余(我自己的网络环境10分钟以内)。
- 一般在有GPU的情况下跑一次需要要多长时间
因为是调用openai的api,没有在本机跑模型,所以GPU不会对结果有任何影响。
- curl指令测试网络结果代表什么意思
可以看到正确返回了结果,所以理论上可以正常访问openai的API
- 调用代码时,网络请求不成功
通过问题3
可以看到,你的环境能正常访问openai的API。
如果你是网络直连(没有使用VPN),那你可以尝试将script/run.py
里的proxy参数设为None
。
如果你使用了VPN,那么你可以尝试将proxy的参数设为你的VPN的http(s)端口地址。参考 #1
from llm4rs.
感谢大佬您的回复!我尝试了您说的方法将proxy设置为None并重新运行,但是仍然不work;之后我按照issue2的方法在post request那里删除了proxy变量,也是仍然不可以。后来我尝试了在openai.py 加了几个print函数,发现queue_of_requests_to_retry.empty() [line 65]在这个过程中始终为False,且下列的代码持续循环,持续print "making next request" 以及 "https://api.openai.com/v1/completions” :
# if enough capacity available, call API
if next_request:
#print("making next request...")
next_request_tokens = next_request.token_consumption
if (
available_request_capacity >= 1
and available_token_capacity >= next_request_tokens
):
# update counters
available_request_capacity -= 1
available_token_capacity -= next_request_tokens
next_request.attempts_left -= 1
# call API
print(request_url)
asyncio.create_task(
next_request.call_API(
request_url=request_url,
request_header=request_header,
retry_queue=queue_of_requests_to_retry,
save_filepath=save_filepath,
status_tracker=status_tracker,
proxy=proxy
)
)
后来我查看了下log日志,发现有两个问题,一个是failed过多(Request 43-0),一个是没有invalid URL, 内容显示如下:
2023-09-13 13:31:16,636 Request {'model': 'text-davinci-003', 'prompt': "You are a movie recommender system now.\nInput: Here is the watching history of a user: Gattaca, Armageddon, Big, Babes in Toyland, Gladiator. Based on this history, please rank the following candidate movies: (A) Con Air (B) Mulan (C) Nikita (D) Donnie Brasco (E) Star Wars: Episode I - The Phantom Menace\nOutput: The answer index is D C B A E.\nInput: Here is the watching history of a user: Sling Blade, Animal House, The Wizard of Oz, Blood Simple, My Life as a Dog. Based on this history, please rank the following candidate movies: (A) Heavenly Creatures (B) Gandhi (C) Diner (D) Antonia's Line (E) A Room with a View\nOutput: The answer index is", 'max_tokens': 20, 'temperature': 0, 'top_p': 1, 'frequency_penalty': 0.0, 'presence_penalty': 0.0, 'stop': '\n', 'logprobs': None, 'logit_bias': {}} failed after all attempts. Saving errors: [, , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , ]
2023-09-13 13:31:16,656 Task exception was never retrieved
future: <Task finished name='Task-3801' coro=<APIRequest.call_API() done, defined at /home/ziyubian/LLM4RS/src/api/openai.py:175> exception=TypeError('Object of type InvalidURL is not JSON serializable')>
Traceback (most recent call last):
File "/home/ziyubian/LLM4RS/src/api/openai.py", line 214, in call_API
append_to_jsonl({"task_id": self.task_id, "target": self.target, "target_index": self.target_index, "pos": self.pos, "request": self.request_json, "result": self.result}, save_filepath)
File "/home/ziyubian/LLM4RS/src/api/openai.py", line 234, in append_to_jsonl
json_string = json.dumps(data)
File "/home/ziyubian/anaconda3/envs/LLM4RS/lib/python3.9/json/init.py", line 231, in dumps
return _default_encoder.encode(obj)
File "/home/ziyubian/anaconda3/envs/LLM4RS/lib/python3.9/json/encoder.py", line 199, in encode
chunks = self.iterencode(o, _one_shot=True)
File "/home/ziyubian/anaconda3/envs/LLM4RS/lib/python3.9/json/encoder.py", line 257, in iterencode
return _iterencode(o, 0)
File "/home/ziyubian/anaconda3/envs/LLM4RS/lib/python3.9/json/encoder.py", line 179, in default
raise TypeError(f'Object of type {o.class.name} '
TypeError: Object of type InvalidURL is not JSON serializable
我是尝试在ubuntu的anaconda虚拟环境下运行这个程序,网络没有问题,可以正常访问openai的API, 想问下大佬还有什么可以用的解决方法吗?麻烦您了
from llm4rs.
抱歉,我目前找不到你那边发生了什么错误,大概率还是网络设置的原因。我重新clone了这个仓库,也用anaconda新建了一个环境。在配置好proxy之后,是能够正常使用的。
from llm4rs.
多谢大佬的及时回复,我再仔细检查一下看看是哪里出了问题,麻烦啦
from llm4rs.
Related Issues (7)
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from llm4rs.