relationship_classifier_multi_retrained

This model is a fine-tuned version of roberta-large on a dataset from OYM 1:1 vs. 2:1 early literacy online tutoring. It achieves the following results on the evaluation set:

• Train Loss: 0.2570
• Validation Loss: 0.3345
• Train Accuracy: 0.8891
• Epoch: 2

Model description

This model is a retrained version of https://huggingface.co/StanfordSCALE/relationship_classifier_multi that corrects for overfitting in the previous model.

Intended uses & limitations

This model was trained on online literacy tutoring data for grades K-2. Generalization beyond this context is uncertain.

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• optimizer: {'name': 'Adam', 'weight_decay': None, 'clipnorm': None, 'global_clipnorm': None, 'clipvalue': None, 'use_ema': False, 'ema_momentum': 0.99, 'ema_overwrite_frequency': None, 'jit_compile': True, 'is_legacy_optimizer': False, 'learning_rate': {'module': 'transformers.optimization_tf', 'class_name': 'WarmUp', 'config': {'initial_learning_rate': 1e-05, 'decay_schedule_fn': {'module': 'keras.optimizers.schedules', 'class_name': 'PolynomialDecay', 'config': {'initial_learning_rate': 1e-05, 'decay_steps': 454, 'end_learning_rate': 0.0, 'power': 1.0, 'cycle': False, 'name': None}, 'registered_name': None}, 'warmup_steps': 50, 'power': 1.0, 'name': None}, 'registered_name': 'WarmUp'}, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-08, 'amsgrad': False}
• training_precision: float32

Training results

Train Loss	Validation Loss	Train Accuracy	Epoch
0.8833	0.6258	0.7781	0
0.4531	0.3678	0.8713	1
0.2570	0.3345	0.8891	2

Framework versions

• Transformers 4.51.3
• TensorFlow 2.19.0
• Datasets 4.8.5
• Tokenizers 0.21.4