How to save the predicted images from inference

[Z0721Z]

I followed the "Inference and Evaluation" tutorial but still couldn't save the prediction results of my dataset.

[Z0721Z]

[hkchengrex]

The outputs should have been saved in the "output" directory.

[Z0721Z]

The outputs should have been saved in the "output" directory.

Thank you for the author's patient response. I have one more question: I would like to know if the main training phase uses a semi-supervised approach, because I noticed that the data loading code seems to require all images in the dataset to have labels.

[hkchengrex]

The training process is not semi-supervised. Semi-supervised VOS refers to the inference process, where the first-frame ground-truth mask is given.