Thank you for your great work. I notice that the model would be updated if I want to predict the bounding box in a new video sequence. However, this seems to be impractical because it might be impossible to retrain the model in the client. Do you have any suggestion for this? Thank you again.
Thank you for your great work. I notice that the model would be updated if I want to predict the bounding box in a new video sequence. However, this seems to be impractical because it might be impossible to retrain the model in the client. Do you have any suggestion for this? Thank you again.