Henghui Ding

> Thank you for your great work! > > I am curious about testing with my own dataset. I just want to get a mask image from my data. >...

Hi, The region queries have different supervision. Split the image to 10x10 patches, then we have 10x10=100 region queries, and a 10x10 mini-map. If the (3, 6)-th image patch contains...

Hi, Thank you for the question. It depends: * The final output mask of LMPM is binary * If the given sentence refers to a single object, then LMPM outputs...

Hi, Thank you very much for your interest. It is evaluated after training with the MeViS training set.

>Hi, > >Thanks for your interest and feedback! We missed it during the data filtering, but this data won't affect the training of existing methods. The validation and testing set...

> Hi, thanks for the data. > > I think there is an issue with the valid dataset on OneDrive/Google Drive. The number of folders in valid dataset is 311,...

Please kindly submit your **val set** results to [**CodaLab online evaluation**](https://codalab.lisn.upsaclay.fr/competitions/15094). For offline validation on you local machine, please use **Valu set** The **test set** is used for challenges, such...

What's your CodaLab username? We cannot identity which one is yours on CodaLad. One possible reason is that you use different file structure from [here](https://github.com/henghuiding/MeViS?tab=readme-ov-file#codalab-evaluation-submission-guideline) required by our CodaLab, e.g.,...

Hi, Thank you for your attention. #4 is not the related issues. The annotation masks of validation split are not released, please submit your validation predictions to [MOSE CodaLab Evaluation...

> When I clicked the 'download data button', it automatically redirected to codelab. But currently I am unable to log in to codlab. > > Could you please send the...