xlm-roberta-large-vieille-france

This model is a fine-tuned version of xaviergillard/xlm-roberta-large-vieille-france on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0099
• Precision: 0.9885
• Recall: 0.9919
• F1: 0.9902

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: 5e-05
• train_batch_size: 32
• eval_batch_size: 32
• 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: 15

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1
No log	1.0	32	0.0061	0.9862	0.9919	0.9891
No log	2.0	64	0.0102	0.9863	0.9931	0.9897
No log	3.0	96	0.0128	0.9795	0.9908	0.9851
No log	4.0	128	0.0106	0.9784	0.9931	0.9857
No log	5.0	160	0.0120	0.9873	0.9896	0.9885
No log	6.0	192	0.0134	0.9817	0.9908	0.9862
No log	7.0	224	0.0097	0.9817	0.9908	0.9862
No log	8.0	256	0.0084	0.9840	0.9942	0.9891
No log	9.0	288	0.0077	0.9885	0.9931	0.9908
No log	10.0	320	0.0101	0.9851	0.9908	0.9879
No log	11.0	352	0.0102	0.9851	0.9919	0.9885
No log	12.0	384	0.0105	0.9862	0.9919	0.9891
No log	13.0	416	0.0102	0.9885	0.9919	0.9902
No log	14.0	448	0.0098	0.9885	0.9919	0.9902
No log	15.0	480	0.0099	0.9885	0.9919	0.9902

Framework versions

• Transformers 4.48.3
• Pytorch 2.1.2
• Datasets 3.3.0
• Tokenizers 0.21.0