anthonymrios / bio-cnn Goto Github PK

Hi,

I tried to run the code with train.py. But it raised the following error:

For the code here: from models.cnn2 import CNN, there is no cnn2.py in the models folder, only cnn.py. You can check the code here:

When I try the cnn.py, it doesn't work, too.

Can you give me any suggestions?

Thanks,

Junjie

Dear authors of bio-CNN,

I was recently trying to reproduce the main paper of this repo (https://www.ncbi.nlm.nih.gov/pubmed/28736769), yet am having a hard time obtaining the train/test datasets, or to be more precise:

1. MED-LINE dataset, containing 89,942 training docs, 500 validation and 46,911 for testing
2. check tags (12), low recall terms (7) and low precision terms (10)
3. Hence, a total of 29 target terms (if I understand correctly, this is still new to me)

Have I simply missed the available datasets, I apologize, yet I cannot seem to find them anywhere.

Thank you very much.