On the partition problem of training sets
Closed this issue · 3 comments
Hello, your training set uses 800 pairs of synthetic datasets, among which there are ten types of water quality in the synthetic dataset. Can you tell me that 800 pairs of datasets extract 80 images from each water quality type? And can the image content extracted from each water quality type overlap or not? (For example, in type I water, if the 18th image is extracted, can the 18th image still be extracted in other types of water?)
Thank you and good luck！
Reproduced the author's paper results, very good work
sorry, may i ask you how to reproduce the paper results score, it seems like the author didn't release the training testing validating datasets ~ thanks
sorry, may i ask you how to reproduce the paper results score, it seems like the author didn't release the training testing validating datasets ~ thanks
Sorry, I just saw the message. I randomly divided the dataset, trained and tested it using the randomly divided dataset.