Laiyi Fu

Hi there, It depends on how you would like to deal with these data. In our work, we merged all these files in the RNAStralign dataset into one folder and...

Hi, We trained/evluated our model with batch size 1. Different sequences have very different running statistics, so we don't want the model to use the previously stored statistics.

Hi there, Sorry for that, since our framework could deal with sequence with various length, so to avoid out-of-memory issue, we have limited the batch size and set it to...

Hi, Regarding to your question on padding batches with different size, I'm afraid we don't have that function. The reason is that different sequence have different length(ranging from 10bp to...

Hi there, Thanks for pointing that out, the config.test line does not have any function in the processed procedure. So you may delete that line to make the script works....

Hi there, Since our model is designed for predicting RNA sequences whose nucleotides are A,U,C,G (instead of A,T,C,G), so you may need to change your sequence to that accordingly. Thanks

Hi, We have updated our web server, that should be work fine right now. Please take a look. Thanks,

Hi there, Thanks for reaching out. The first issue you point out is due to the GPU id, you can switch to your own cuda device id or change it...

Hi, I guess it's because you only run the code on CPU. In that case, you can alter the code: https://github.com/uci-cbcl/UFold/blob/2b5fab25748c2ec6ce39d5b7e5d32a3c78b850de/ufold_predict.py#L297 to: device = torch.device("cpu") to avoid this issue. Please...

Hi there, I believe we run for about 100 epochs to get the score. As for other datasets, we can not guarantee all the epochs are the same because sometimes...