How to segment my own datasets?

[glode515]

trafficstars

How can I segment my image or video?

[hkchengrex]

Does this help: https://github.com/hkchengrex/XMem/blob/main/docs/INFERENCE.md?

[glode515]

I found eval.py can only infer specific datasets, such as "16/D17/Y18/Y19/LV1/LV3/G"

[hkchengrex]

Did you see the "On custom data" section?

[mjw123bs]

If I want to train my own dataset, do I need to set up the datasets in it according to GETTING_STARTED.md?

[hkchengrex]

@mjw123bs If you are not going to use them then no.

[mjw123bs]

@mjw123bs If you are not going to use them then no.

ok thanks, i'll try it.

[mjw123bs]

@mjw123bs If you are not going to use them then no.

Hi, I have some problems, how to train on custom data with a single gpu? And in the "on custom data" section, is .png a mask? Is video1 the original video?

[hkchengrex]

@mjw123bs

1. I don't know how is your data formatted. I imagine you can copy/paste/edit some of our loading code in train.py and dataset/vos_dataset.py.
2. Yes. Like DAVIS/YouTubeVOS.

[mjw123bs]

Thanks for the quick reply！ I'll try it.

[Eliza-wxq]

Thanks for the quick reply！ I'll try it.

Could you please tell me whether you have successfully deployed? Train your own data set on a single GPU.

[mjw123bs]

Thanks for the quick reply！ I'll try it.

Could you please tell me whether you have successfully deployed? Train your own data set on a single GPU.

Sorry, no, I haven't done any research since then. I will inform you as soon as I get the results.

[Eliza-wxq]

Thank you very much for your quick reply