Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Image Generation with ChatGpt Plus #2453

Open
judo-sudo opened this issue Dec 4, 2024 · 6 comments
Open

Image Generation with ChatGpt Plus #2453

judo-sudo opened this issue Dec 4, 2024 · 6 comments
Assignees
Labels
bug Something isn't working

Comments

@judo-sudo
Copy link

i am trying to generate image with a chatgpt plus (paid) account.

        response = client.chat.completions.create(
            messages=[{"role": "user","content": promptText}] ,
            model='auto',
            conversation_id="674fef59-d9e8-800f-9621-xxxxxxxx",
            history_disabled=False,
            auth=True,
            provider=g4f.Provider.OpenaiChat)

The conversation id belongs to a conversation with dall-e, yet I get get an error saying: "I currently can't generate images, but I can describe how you might create one.... etc"

setting model to 'gpt-4o' does not work either.

The same code used to work until 3 weeks ago.

I am using g4f 0.3.8.1

@judo-sudo judo-sudo added the bug Something isn't working label Dec 4, 2024
@judo-sudo
Copy link
Author

the bug seems that, irrespective of what model (auto or dall-e-3) you specify in the model parameter,
the request seems to be sent to gpt-4o-mini; which is not capable of generating images.

same bug in client.images.generate too.

@hlohaus
Copy link
Collaborator

hlohaus commented Dec 4, 2024

I can't find anything wrong.

Screenshot_20241204_222813_Chrome.jpg

@judo-sudo
Copy link
Author

judo-sudo commented Dec 5, 2024

Here are my logs:

It seems to send the request to wrong model, even if I select the dall-e model for image generation.

Screenshot 2024-12-04 at 8 20 52 PM

Using OpenaiChat provider and dall-e-3 model
Arkose: False
Proofofwork: gAAAAABWzMxM...
AccessToken: eyJhbGciOiJS...
OpenaiChat: New conversation: 

@hlohaus
Copy link
Collaborator

hlohaus commented Dec 5, 2024

If you select dall-e-3 will be used "auto". Try choosing a different model.

@judo-sudo
Copy link
Author

If i choose another model (like 4o) , I am not able to generate images. I get an error saying "I cannot generate images..." or "Image generation is currently unavailable"....

i get the same error when I try the 4o model via chatgpt.com.

However, I can generate images if I choose dall-e on chatgpt.com.

If there were were a way to force use of dall-e via g4f.

I also tried the following code, but no avail:


        response = client.chat.completions.create(
            messages=[{"role": "user","content": promptText}] ,
            model='dall-e-3',
            conversation_id="a-dalle-conversation-id",
            history_disabled=False,
            auth=True,
            provider=g4f.Provider.OpenaiChat)

@hlohaus
Copy link
Collaborator

hlohaus commented Dec 10, 2024

ChatGPT's image generation incorporates robust bot protection. Using outdated or self-generated proof tokens will prevent image generation; however, utilizing the current version resolves this issue.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
bug Something isn't working
Projects
None yet
Development

No branches or pull requests

3 participants