This is a fine-tuned BART model for Definition Generation. It is still in the prototype stage, fine-tuned only with 40k Training Instances of (definition, context) pairs for 3 epochs. The eval_loss is still in 2.30. The beam Size is 4.
This is a fine-tuned BART model for Definition Generation. It is still in the prototype stage, fine-tuned only with 40k Training Instances of (definition, context) pairs for 3 epochs. The eval_loss is still in 2.30. The beam Size is 4.