Question about training

[AlbertoN3to]

Hi!

I wanted to know if is there a way to train with different unannotated datasets (using the pretrained model) from different sign languages.
From a glance, doesn't seem like the code supports this.

Thanks in advance :)

[RenzKa]

Hi,

Sorry for the late reply.
Yes this does not work out-of-the-box since it depends a little bit in what format your dataset is.

I think the easiest way is to look at the Dataloader for our datasets (folder /datasets) and adapt this for a Dataloader for your data.

Best,
Katrin