hackathon-pln-es/roberta-base-bne-squad2-es

roberta-base es for QA

This model is a fine-tuned version of PlanTL-GOB-ES/roberta-base-bne on the squad_es(v2) training dataset.

Hyperparameters

The hyperparameters were chosen based on those used in deepset/roberta-base-squad2, an english-based model trained for similar purposes

 --num_train_epochs 2
 --learning_rate  3e-5
 --max_seq_length 386
 --doc_stride 128 

Performance

Evaluated on the squad_es(v2) dev set.

  eval_exact": 62.13526733007252,
  eval_f1": 69.38515019522332,  
  
  
  eval_HasAns_exact": 53.07017543859649,
  eval_HasAns_f1": 67.57238714827123,
  eval_HasAns_total": 5928,
  eval_NoAns_exact": 71.19730185497471,
  eval_NoAns_f1": 71.19730185497471,
  eval_NoAns_total": 5930,
  

Team

Santiago Maximo: smaximo