search

abi / screenshot-to-code

Drop in a screenshot and convert it to clean code (HTML/Tailwind/React/Vue)
https://screenshottocode.com
MIT License
60.52k stars 7.43k forks source link

How to switch models #413

Open EZzer0 opened 1 month ago

EZzer0 commented 1 month ago

There is no way to find a button to switch models, for example, in a video, and the Llm.CLAUDE_3_5_SONNET_2024_06_20 is called by default every time

abi commented 1 month ago

Hi there, right now, there isn't a UI to switch models. For images, we use GPT4o and Claude 3.5 Sonnet, I believe Option 1 is Claude 3.5 Sonnet and the other is GPT4o if you have both API keys set. If you only have one API key set, then we just generate two options with the same model.

Which model are you looking to use for video? I think video always uses Claude 3 Opus.

EZzer0 commented 1 month ago

image image "As you can see in the screenshot, I used a third-party API and changed the base URL, but my API only supports models like GPT-4 and doesn't support CLAUDE_3_5. Because of this, no code was generated. I tried to tweak the backend, but I couldn't find a fix, which caused the error: [HTML Extraction] No tags found in the generated content."

abi commented 1 month ago

If you look like the line that says models used, it says OpenAI twice. Is any code being returned if you look at the code tab?

Does your third party API work for other use cases?

This likely is an issue with your API not this repo.