claims-detection-model-v1

This model is a fine-tuned version of dccuchile/bert-base-spanish-wwm-uncased on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0885
• Accuracy: 0.8836
• Precision: 0.8275
• Recall: 0.9088
• F1: 0.8662
• F1 Macro: 0.8816

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: 8.182441148887555e-05
• train_batch_size: 32
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 64
• 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
• lr_scheduler_warmup_steps: 0.05924145688620425
• num_epochs: 4

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1	F1 Macro
0.2216	0.5814	100	0.0920	0.84	0.7897	0.8368	0.8126	0.8365
0.1603	1.1628	200	0.0836	0.8727	0.8120	0.9018	0.8545	0.8707
0.1409	1.7442	300	0.0820	0.8756	0.8223	0.8930	0.8562	0.8733
0.1319	2.3256	400	0.0812	0.8764	0.8096	0.9175	0.8602	0.8747
0.1094	2.9070	500	0.0832	0.8807	0.8192	0.9140	0.8640	0.8789
0.0909	3.4884	600	0.0885	0.8836	0.8275	0.9088	0.8662	0.8816

Framework versions

• Transformers 5.0.0
• Pytorch 2.10.0+cu128
• Datasets 4.0.0
• Tokenizers 0.22.2