nwams / clinicalbert-deep-learning--predicting-hospital-readmission-using-transformer Goto Github PK

Hi there,

The link for pretrain model at Kexin's Github is expired, and I am wondering if you happen to have a copy of it.

Thanks a lot!

Hi,

I found we cannot download pre-trained models. Can you upload these again?

Thanks for your awesome work!
I wonder if you tested trained weights on texts that are not in the dataset. Cause I got quite low probability (0.17) for a patient that is old, has chronic diseases and has high chance of readmission. And at the same time, a young men admitting for the first time with cuts has prob 0.34. Have you seen such kind of model behaviour?

Thank you for implementing ClinicalBERT on Transformer. It looks the google drive link you shared for the pre-trained model is no longer working. Could you please check it?

"You can use this google link to download the pre-trained ClinicalBERT model along with the readmission task fine-tuned model weights."