matthewleechen/time-saving_stated_aim_classifier

Time Saving Stated Aim Classifier

This is a roberta-base model that is trained to classify whether a set of explicit stated aims extracted from a British historical patent includes a time-saving objective.

Labels were manually generated and then checked with Gemini 2.0 Flash with the attached prompt.

Hyperparameters: lr = 3e-5 batch size = 128

Test set results:

{'eval_loss': 0.5350350737571716,
 'eval_accuracy': 0.89,
 'eval_precision': 0.8972633319553535,
 'eval_recall': 0.89,
 'eval_f1': 0.8891426257718393,
 'eval_runtime': 0.4081,
 'eval_samples_per_second': 245.053,
 'eval_steps_per_second': 2.451}