Training on Custom dataset

[Aayushktyagi]

Hi,

Can you provide steps to train on custom dataset? I have dataset in coco format. I have registered the data and it's running fine for now.

Can you point where one has to change the num_classes =1? I am getting warning like:

WARNING [02/11 12:56:25 fvcore.common.checkpoint]: Skip loading parameter 'teacher.proposal_generator.fcos_head.cls_logits.weight' to the model due to incompatible shapes: (80, 256, 3, 3) in the checkpoint but (1, 256, 3, 3) in the model! You might want to double check if this is expected. WARNING [02/11 12:56:25 fvcore.common.checkpoint]: Skip loading parameter 'teacher.proposal_generator.fcos_head.cls_logits.bias' to the model due to incompatible shapes: (80,) in the checkpoint but (1,) in the model! You might want to double check if this is expected. WARNING [02/11 12:56:25 fvcore.common.checkpoint]: Skip loading parameter 'student.proposal_generator.fcos_head.cls_logits.weight' to the model due to incompatible shapes: (80, 256, 3, 3) in the checkpoint but (1, 256, 3, 3) in the model! You might want to double check if this is expected. WARNING [02/11 12:56:25 fvcore.common.checkpoint]: Skip loading parameter 'student.proposal_generator.fcos_head.cls_logits.bias' to the model due to incompatible shapes: (80,) in the checkpoint but (1,) in the model! You might want to double check if this is expected.

Is it fine?

Thanks