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

Null Action but the console output shows the correct(?) action coordinates #37

Open
Jake36921 opened this issue Jan 24, 2025 · 2 comments

Comments

@Jake36921
Copy link

2025-01-24 14:13:17 [INFO]
[LM STUDIO SERVER] [ui tars] Generated prediction: {
"id": "chatcmpl-b2m6rawskj8wm4x4ca980m",
"object": "chat.completion",
"created": 1737699167,
"model": "ui tars 7B q8.gguf",
"choices": [
{
"index": 0,
"logprobs": null,
"finish_reason": "stop",
"message": {
"role": "assistant",
"content": "1. The user has already typed "youtube.com" into the address bar, which is the first step in accessing the YouTube website. The next logical action is to press the Enter key to submit the input and load the website.\n2. Pressing Enter will confirm the URL and prompt the browser to navigate to YouTube, completing the task of opening the site.\nAction: click(start_box='<|box_start|>(691,85)<|box_end|>')"
}
}
],
"usage": {
"prompt_tokens": 443,
"completion_tokens": 89,
"total_tokens": 532
},
"system_fingerprint": "ui tars"
}

Image

@mingcomplex
Copy link
Collaborator

mingcomplex commented Jan 24, 2025

Hello, are you using the GGUF model?

@mingcomplex
Copy link
Collaborator

mingcomplex commented Jan 24, 2025

The GGUF model has some issues with performance due to quantization, and the output might not follow the expected format. Please use the full-size model instead.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants