Hi,
I run the code published in this repo, but i didn't get the same performance as mentioned in Performance.txtfile.
i just copied and pasted the code and corrected some typos like putting ( ) for print function and just made it run.
i used Colab Pro in the implementation and the results of DRIVE dataset is as follow:
Area under the ROC curve: 0.9745132983506472
Area under Precision-Recall curve: 0.9005084432534444
Global Accuracy: 0.9542264313839384
Specificity: 0.9833227875108509
Sensitivity: 0.7547351419287491
Precision: 0.8684320607463418
Jaccard similarity score: 0.6772917155378523
F1 score (F-measure): 0.807601574924213
Also, i tried to load the test_best_weights.h5 to see the prediction's of your trained model (without retraining from scratch), but i also get the below error while reading test_architecture.json file, so i couldn't even load your trained model and evaluated.
Hi, I run the code published in this repo, but i didn't get the same performance as mentioned in
Performance.txt
file. i just copied and pasted the code and corrected some typos like putting ( ) for print function and just made it run. i used Colab Pro in the implementation and the results of DRIVE dataset is as follow:Also, i tried to load the
test_best_weights.h5
to see the prediction's of your trained model (without retraining from scratch), but i also get the below error while readingtest_architecture.json
file, so i couldn't even load your trained model and evaluated.ValueError: ('Unrecognized keyword arguments:', dict_keys(['input_dtype']))
Please, what should i do to get the same performance as yours ? Any help is appreciated. Thanks