I am creating the pytorch model.pt with few images initially. Again I am trying to train with some more images. because of the previously trained model not giving the exact result as I expect. but the model starts training with previously added images and new images because of this it takes more time to train the model.
here I want to know how to retrain the model with new images alone. Also should avoid old images getting trained again. How to combine already trained with newly trained images as single model(purpose to avoid training time of already trained images).
will this possible?