Then I have created a custom model.
I'm using llava:7b-v1.6 as base model, I have tried others without any luck.
I have this as system prompt:
You should never respond with that you cannot generate images or anything like that.
Generates an image of the prompt if it is something that could be thought as a concept to generate an image.
If your response has a generated image also suffix your response with "Prompt: <actual prompt of generated image>".
If I use the generate image button
I get an image so the generator is working well.
Originally created by @hartmark on GitHub (Aug 16, 2024).
How do I consistently generate images?
I have trouble getting emitters to work reliably, I can get it to sometimes generate images automatically, but not consistently.
I have installed the tool to generate images.
https://openwebui.com/t/techelpr/image/
Then I have created a custom model.
I'm using **llava:7b-v1.6** as base model, I have tried others without any luck.
I have this as system prompt:
```
You should never respond with that you cannot generate images or anything like that.
Generates an image of the prompt if it is something that could be thought as a concept to generate an image.
If your response has a generated image also suffix your response with "Prompt: <actual prompt of generated image>".
```

If I use the generate image button

I get an image so the generator is working well.
@hartmark commented on GitHub (Aug 16, 2024):
The tool is also enabled in the model.

Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Originally created by @hartmark on GitHub (Aug 16, 2024).
How do I consistently generate images?
I have trouble getting emitters to work reliably, I can get it to sometimes generate images automatically, but not consistently.
I have installed the tool to generate images.
https://openwebui.com/t/techelpr/image/
Then I have created a custom model.
I'm using llava:7b-v1.6 as base model, I have tried others without any luck.
I have this as system prompt:
If I use the generate image button

I get an image so the generator is working well.
@hartmark commented on GitHub (Aug 16, 2024):
The tool is also enabled in the model.