Bi-SRNet
Bi-SRNet copied to clipboard
ggsDing
other dataset
Hello, I tried to test your proposed model on the Landsat-SCD dataset, but it didn't work well, I tried making the following changes: num_classes = 5 ST_COLORMAP = [[255,255,255], [128,128,128], [130,87,87], [255,0,0], [0,0,255]] ST_CLASSES = ['unchanged', 'farmland', 'desert', 'building', 'water'] The above modifications occur in : ./ datasets/RS_ST.py. I don't know if you have done experiments on this dataset, can you give me a little comment?
Hello, I followed your part of the work, but the results of my training with Bi-SRNet are not very good, mIoU is only 81, did you make other parameter adjustments?
In the published paper the highest mIoU is 73.41. Did you do experiments with your own settings?
In the published paper the highest mIoU is 73.41. Did you do experiments with your own settings?
Maybe I didn't make it clear what I meant, but I was talking about experiments on the Landsat dataset.
In the published paper the highest mIoU is 73.41. Did you do experiments with your own settings?
Maybe I didn't make it clear what I meant, but I was talking about experiments on the Landsat dataset.
In the dataset mentioned in another work you gave, I found that the results of your training are very good, and I would like to ask if you have made some parameter adjustments.
In the published paper the highest mIoU is 73.41. Did you do experiments with your own settings?
Maybe I didn't make it clear what I meant, but I was talking about experiments on the Landsat dataset.
In the dataset mentioned in another work you gave, I found that the results of your training are very good, and I would like to ask if you have made some parameter adjustments.
Please discuss under the SCanNet repository. The code will be made available soon.