ParikhKadam / bidaf-keras

Bidirectional Attention Flow for Machine Comprehension implemented in Keras 2
GNU General Public License v3.0
64 stars 21 forks source link

How 'bidaf_50.h5' is trained? #20

Open bbaa2837 opened 4 years ago

bbaa2837 commented 4 years ago

Hello. Thanks for your work. By the way, I'm wondering in which environment did you gained (e.g. the number of epochs) the pre-trained model 'bidaf_50.h5'. I'll really appreciate if you can elaborate in detail.

Thanks

ParikhKadam commented 4 years ago

Hey @bbaa2837

_50 here means the model trained in 50 epochs. I might have also provided a model _10 which is the model trained till 10 epochs. The parameters for training are default one provided in the main file.