ruu3f/freegpt

freeGPT provides free access to text and image generation models.

Hi, I'm Ruu3f! 👋

I'm a full-stack developer with a passion for open-source projects. If you have any questions or need help regarding my projects, feel free to reach out to me on Discord.

Skills

Stats

freegpt's Issues

gpt4.py script does no longer work

gpt4.py script does no longer work , It replies with "Unable to fetch response."

Can't use falcon_40b or alpaca_7b

Describe the bug
If you try to use falcon_40b or alpaca_7b, whether in synchro or in ansychromode it doesn't work. The Error I get is 'conversationID'

To Reproduce
Paste this code:
from freeGPT import AsyncClient
from asyncio import run

async def main():
while True:
prompt = input("👦: ")
try:
resp = await AsyncClient.create_completion("falcon_40b", prompt)
print(f"🤖: {resp}")
except Exception as e:
print(f"🤖: {e}")

run(main())

or paste this code:

from freeGPT import Client

while True:
prompt = input("👦: ")
try:
resp = Client.create_completion("falcon_40b", prompt)
print(f"🤖: {resp}")
except Exception as e:
print(f"🤖: {e}")

Expected behavior
A clear and concise description of what you expected to happen.

Desktop:

• OS: Windows 11 (doesn't work on Google Collab either)

Additional context
I tried many things to fix this "bug", but I can't find the solution for this problem, except using gpt3 or gpt4, but for my Ai-project its very important to use falcon_40b.

Chat History Request

Hey developer thanks for creating such a great repo for us but I have a question is it possible to send a request along with chat history (in gpt4 and alpaca_7b )? if yes then how?

GPT4 isnt working

bots does not have memory

why do gpt3 or gpt4 bots not have memory? Although they have memory on their websites whether you.com or AIChatOS .
Appreciate your response

Can you add new provider?

chat.ylokh.xyz this provider have a lot models include GPT 4 for free and have params to set temperatur,top_p and etc.

GPT3_5 not working since yesterday

👦:
🤖: 'choices'

Are you paying for this or what

Im kinda interested

gpt3 not working

I use the gpt3 model, and about a week ago it started to act strangely. This is my code:

from freeGPT import Client as Client2

response = await asyncio.to_thread(Client2.create_completion, "gpt3", prompt)

I get this output:

📢 通知
接上级主管部门通知，网站：[https://c2.binjie.fun](https://c2.binjie.fun) 已于2023年10月24日晚上18:00前完全下架AI聊天对话功能。后续上线工作需等待网信办工作人员审核和评估，敬请谅解。
🔗 用户可前往 https://chat18.aichatos8.com/ 继续使用（这个能用的网站10月24号即明天中午12点后不再显示，可立即收藏）
🔗 本网站无任何微信公众号 qq群等其他平台社群, 望用户能及时收藏本站, 获得最新消息请关注网站新聊天页面的公告!!!

This translates to:

📢 Notice
According to the notice from the superior competent department, the website: [https://c2.binjie.fun](https://c2.binjie.fun) has completely removed the AI ​​chat function before 18:00 on October 24, 2023. The subsequent online work needs to wait for the review and evaluation of the staff of the Cyberspace Administration of China, please understand.
🔗 Users can go to https://chat18.aichatos8.com/ to continue using it (this usable website will no longer be displayed after 12 noon on October 24, and can be bookmarked immediately)
🔗 This website does not have any WeChat public account, QQ group or other platform communities. I hope users can bookmark this site in time. Please pay attention to the announcement of the new chat page on the website to get the latest news!!!

Failed to perform, ErrCode: 77

Describe the bug
A clear and concise description of what the bug is.
when i converted 'py' file to .exe i faced this issue
(Failed to perform, ErrCode: 77, Reason: 'error setting certificate verify locations: CAfile: C:\Users\JooDe\AppData\Local\Temp_MEI99322\curl_cffi\cacert.pem CApath: none')
i don't know why its brings, when its already working in 'py' file without errors.

• OS: Windows 10 64bit
• Python Version: 3.11.2

requirements.txt

I'm trying to install this on a Raspberry Pi, but it won't run because of:

ModuleNotFoundError: No module named 'curl_cffi'

What about adding a requirements.txt to the repo?

Thanks!

403 error on gpt3

Hi ruu3f
Please solve this Bug

BUG
403 error on gpt3

Code
import freeGPT
from asyncio import run

async def main():
while True:
prompt = input("👦: ")
try:
resp = await getattr(freeGPT, "gpt3").Completion().create(prompt)
print(f"🤖: {resp}")
except Exception as e:
print(f"🤖: {e}")

run(main())

Output
PS C:\Users\HP> python -u "c:\Python_files\main.py"
👦: hi
🤖:
👦: who are you
🤖:

Statement
You can see i asked hi and tell nothing
So i saw the source code
And i add extra one return value to find problem
old code = return resp
changed code = return resp,resp_obj

run that code after changing

output
👦: hi
🤖: ('', <ClientResponse(https://ava-alpha-api.codelink.io/api/chat) [403 Forbidden]>
<CIMultiDictProxy('Date': 'Sun, 27 Aug 2023 17:23:00 GMT', 'Content-Type': 'text/html', 'Transfer-Encoding': 'chunked', 'Connection': 'keep-alive', 'x-powered-by': 'Express', 'Access-Control-Allow-Origin': '*', 'x-cloud-trace-context': 'dcdfc9a5c798b7ae7b669395f77c257a', 'Via': '1.1 google', 'CF-Cache-Status': 'DYNAMIC', 'Report-To': '{"endpoints":[{"url":"https:\/\/a.nel.cloudflare.com\/report\/v3?s=mZGmHaoXnCDP4McZ7JWbS6js9G63OsAUbjicRUWbBgeqG4%2FNdOqZ1DOjRSL3yiqlTwOA3g1F0uywRD6AfjpnMJe7yq866ZaJL3JZhU2QV45yQDzTNTNTIlqNXCeGL0XFZHEBnCZoIlwWTo3Tutg8t0yHCYRZgZA%3D"}],"group":"cf-nel","max_age":604800}', 'NEL': '{"success_fraction":0,"report_to":"cf-nel","max_age":604800}', 'Server': 'cloudflare', 'CF-RAY': '7fd5fff4eba947c0-BOM', 'Content-Encoding': 'gzip')>
)

conclusion
i found your code have 403 request error
Please debug this problem

not actually gpt-4

part of the supposed gpt-4's response was "model":"gpt-3.5-turbo-0301"
can you find another backend?

Both gpt3 and gpt4 does not work!

Trying the code in the readme file for gpt3 throws session not found error. while trying gpt4 throws a bunch of json data and just crashes.

why tls_cilent

Do you think that both streamlit and hugging face restricted their servers not to scrap data from websites ? I think so because if the issue is in tls_client ,it will not work locally, but since it worked then the issue is in their servers What is your opinion?

Also when I converted my web app to executable file also it didn't fetch the response , I am astonishing regarding this? Any contribution will be appreciated.

The last one, why did you use tls_client not requests or aiohttp like in your gpt3 script?

while i am using the gp4 model in the below code

from freeGPT import Client

while True:
prompt = input("👦: ")
try:
resp = Client.create_completion("gpt4", prompt)
print(f"🤖: {resp}")
except Exception as e:
print(f"🤖: {e}")
in the above code i am using the gpt4 it is generating the output as "🤖: Unable to fetch the response."

hosting-related issue

I built an app depends on your gpt4.py script and it works properly locally, however, when I hosted my app on streamlit and Huggingface spaces both of them replied with the Exception ("Unable to fetch response.") . Note that both servers accept gpt3.py script and fetch the answer using it without problems. I asked ChatGPT and gave him the code of both scripts and he replied that the issue may be with the tls_client library
what is the reason of this ? and how can i solve it?

gpt3 accepts very small number of tokens

Why gpt3 accepts very small nubmer of tokens (around 200 tokens) and when I provides more than this number it returns zero-length answer?

gpt4: Unable to fetch response

Describe the bug
When using the "gpt4" model, the package cannot fetch a response.

To Reproduce
Steps to reproduce the behavior:

1. Copy one of the examples at the initial page.
2. Change the line which initializes the resp variable to use gpt4 as a model).
3. Run the script.
4. Type anything to submit to the model, hit enter.
5. See error

Expected behavior
The script should display the model's response to the prompt

Screenshots

Desktop (please complete the following information):

• OS: Ubuntu 22.04 (but it happens on Macos as well)

Upon installing the FreeGPT library, I encountered the following error. Could you please assist me in identifying the issue? Thank you.

Use `Search Web (Experimental)` feature of H2O.AI

Is your feature request related to a problem? Please describe.
Add a feature so we can use the Search Web (Experimental) feature of H2O.AI (https://gpt-gm.h2o.ai/) model Falcon to access internet option

Describe the solution you'd like
You see this thing? yeah

name 'Session' is not defined

Hi! I'm facing an error when trying to use gpt3

from freeGPT import gpt3

try:

Remove the 'proxy' variable and the 'proxies' parameter if you don't want to use a proxy.

resp = gpt3.Completion.create(prompt=prompt)
print(f"🤖 > {str(resp['text'])}")
except Exception as e:
print(f"🤖 > {str(e)}")

Output:
🤖 > name 'Session' is not defined