Is the model in the web interface different than the one released in the repository?

[cuevhv]

Hi. I was testing the code and I could notice that the model downloaded from the repository performs slightly worse than the one in your website interface.

For example in the attached image. The left one is the output of the website version, and the right one is the results using the code provided in the README. Both have the prompt "person".

Could you please tell if this is due to a difference in the model or a different post-processing.

Thank you!

[haitxabigames]

Did U Use their best model ?

[cuevhv]

@haitxabigames I couldn't find a way to choose the model. How can I choose their best model after download the weights from huggingface?
I use video_predictor = build_sam3_video_predictor() to instantiate the video predictor.

Thanks for he help!