Pretrained_model

[rishabh279]

Can you please provide pretrained_model for DISO as you have provided for PRGE.

[JohnGiorgi]

Yes! I will try to have that done before the end of the week.

[rishabh279]

Thank you!

[JohnGiorgi]

Training... ⏲

[JohnGiorgi]

Okay, the model has be added. You can play around with it in a jupyter notebook with:

from saber import Saber()
saber = Saber()
saber.load('DISO')
text = '''Chronic kidney disease is a worldwide public health problem with an increasing incidence and prevalence, poor outcomes, and high cost. Outcomes of chronic kidney disease include not only kidney failure but also complications of decreased kidney function and cardiovascular disease. Current evidence suggests that some of these adverse outcomes can be prevented or delayed by early detection and treatment. Unfortunately, chronic kidney disease is underdiagnosed and undertreated, in part as a result of lack of agreement on a definition and classification of its stages of progression.'''
saber.annotate(text, jupyter=True)

which gives:

or use the local web-service.

Just a few things to note:

• I didn't do any serious testing of its performance, but I did hit it with some abstracts and it appears to perform well.
• To keep the model small, I loaded only pre-trained word embeddings that were contained in the training sets. You would likely get better performance by loading all available pre-trained word embeddings, but this makes the model huge (~4GB). If I can find a place to host the models I will stick the full-size ones there (maybe google drive).

Please let me know if that works!

[rishabh279]

Sure and thanks a lot!! . I will try the model and let you know.