How to set a proper threshold?

[ayiyayi]

Hello!

Thank you for your great work and for sharing the code!

I would like to ask: how should I set a proper threshold for generating captioning data? Are there recommended values or best practices for choosing this threshold?

Thank you for your help!

[Cooperx521]

Threshold = 0 is good for generating caption. You can set the threshold to -1 if you want to apply stricter filtering for hallucinations.

Best,
Long