vit-ena24-clase

This model is a fine-tuned version of google/vit-base-patch16-224-in21k on the ena24_MD dataset. It achieves the following results on the evaluation set:

• Loss: 0.3132
• Accuracy: 0.9321
• F1: 0.8789

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.0002
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 1
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
1.9153	0.1302	100	1.7647	0.5725	0.4646
1.2463	0.2604	200	1.1008	0.7641	0.6933
0.884	0.3906	300	0.9143	0.7832	0.7113
0.6852	0.5208	400	0.6161	0.8649	0.8027
0.5318	0.6510	500	0.4691	0.8947	0.8376
0.5544	0.7812	600	0.3783	0.9153	0.8984
0.2321	0.9115	700	0.3132	0.9321	0.8789

Framework versions

• Transformers 4.55.0
• Pytorch 2.6.0+cu124
• Datasets 4.0.0
• Tokenizers 0.21.4