Could you please provide the three seeds that you use for random experimental runs in table I?

Jiaoyang45 / Rolling-Unet

MIT License

32 stars 4 forks source link

Could you please provide the three seeds that you use for random experimental runs in table I? #4

Open XGGNet opened 2 months ago

XGGNet commented 2 months ago

Could you please provide the three seeds that you use for random experimental runs in table I? So that We can inherit the results your report on the performance of other working experiments.

Jiaoyang45 commented 2 months ago

In the paper, for the BUSI, GlaS, and CHASEDB1 datasets, we performed three random splits using the train_test_split function, with corresponding random seeds of 41, 1029, and 3407, respectively. ISIC 2018 dataset only performed random segmentation once, with a corresponding seed of 41. To ensure the reproducibility of the experiment, we strictly set the random seeds for functions such as Torch, Python, and Numpy to 1029 using the seed_torch function.