superhero-distilbert-predictor

This model is a fine-tuned version of distilbert-base-uncased on the superhero-texts dataset. This model maps brief descriptions of popular superheroes to their respective comic book universes.

It achieves the following results on the evaluation set:

• Loss: 0.0161
• Accuracy: 1.0
• F1: 1.0
• Precision: 1.0
• Recall: 1.0

Intended uses & limitations

This model is strictly intended for educational use. Do not use this model to draw real world conclusions.

Training and evaluation data

This model was trained on an augmented set of 1100 synthetically generated superhero descriptions and their respective universe label. This model was validated against a set of 100 original, human curated descriptions.

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 2e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall
0.4148	1.0	88	0.2780	0.9489	0.9299	0.9127	0.9489
0.0861	2.0	176	0.0616	0.9830	0.9771	0.9721	0.9830
0.0227	3.0	264	0.0174	1.0	1.0	1.0	1.0
0.0118	4.0	352	0.0099	1.0	1.0	1.0	1.0
0.0074	5.0	440	0.0088	1.0	1.0	1.0	1.0

Framework versions

• Transformers 4.56.1
• Pytorch 2.8.0+cu126
• Datasets 4.0.0
• Tokenizers 0.22.0