Open nihirv opened 9 months ago
Hi, is there any way to pass a batch_size param to either the GroundedSAM constructor or model.label()? Training one-by-one is a bit slow and doesn't utilise my GPU memory is efficiently as it could
model.label()
We do not support this at the moment. If you would like to submit a PR, we can review it!
Hi, is there any way to pass a batch_size param to either the GroundedSAM constructor or
model.label()
? Training one-by-one is a bit slow and doesn't utilise my GPU memory is efficiently as it could