Billsum Evaluation
#6
by
mlkorra
- opened
Have you fine-tuned your model on Billsum data or used the model directly to get the evaluation results (rogue scores) ?
Hey! the model was not fine-tuned on any billsum data, the eval was added as a "test of generalization" as I find "pre-fine-tuning" on booksum enables generally good performance on other tasks and or use cases, see the long-t5-base I trained for example
pszemraj
changed discussion status to
closed