xlm-roberta-large-clinical-ner-vocabulario-extendido-alingRau-sp

This model is a fine-tuned version of anvorja/xlm-roberta-medico-vocabulario-extendido on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0173
• Precision: 0.9817
• Recall: 0.9889
• F1: 0.9853
• Accuracy: 0.9951

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: 1e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 64
• 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: cosine_with_restarts
• lr_scheduler_warmup_ratio: 0.2
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
2.4547	1.0	86	2.2205	0.0	0.0	0.0	0.6584
1.1284	2.0	172	1.0474	0.4300	0.3110	0.3610	0.7759
0.5552	3.0	258	0.4240	0.7206	0.7343	0.7274	0.9076
0.2985	4.0	344	0.2035	0.8502	0.8677	0.8589	0.9508
0.1927	5.0	430	0.1326	0.8683	0.9035	0.8856	0.9658
0.1295	6.0	516	0.1005	0.9052	0.9309	0.9179	0.9747
0.1145	7.0	602	0.0898	0.9151	0.9483	0.9314	0.9776
0.106	8.0	688	0.0619	0.9446	0.9536	0.9491	0.9846
0.0749	9.0	774	0.0509	0.9548	0.9678	0.9613	0.9870
0.0764	10.0	860	0.0441	0.9601	0.9758	0.9678	0.9895
0.0485	11.0	946	0.0375	0.9652	0.9800	0.9725	0.9910
0.0431	12.0	1032	0.0275	0.9698	0.9831	0.9764	0.9927
0.0485	13.0	1118	0.0309	0.9713	0.9815	0.9764	0.9926
0.037	14.0	1204	0.0229	0.9775	0.9863	0.9819	0.9942
0.045	15.0	1290	0.0200	0.9796	0.9863	0.9829	0.9947
0.0286	16.0	1376	0.0173	0.9817	0.9889	0.9853	0.9951
0.0348	17.0	1462	0.0177	0.9780	0.9847	0.9814	0.9945
0.0299	18.0	1548	0.0167	0.9796	0.9858	0.9827	0.9948
0.0308	19.0	1634	0.0165	0.9811	0.9868	0.9840	0.9950
0.0301	19.7719	1700	0.0164	0.9806	0.9868	0.9837	0.9951

Framework versions

• Transformers 4.49.0
• Pytorch 2.6.0+cu124
• Datasets 3.4.1
• Tokenizers 0.21.1