xlm-roberta-large-vieille-france-v2

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: 0.0960
• Precision: 0.7626
• Recall: 0.8061
• F1: 0.7838

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	54	0.0609	0.7207	0.7966	0.7568
No log	2.0	108	0.0541	0.7140	0.8054	0.7570
No log	3.0	162	0.0585	0.7595	0.8083	0.7831
No log	4.0	216	0.0592	0.7384	0.8361	0.7842
No log	5.0	270	0.0684	0.7379	0.7827	0.7597
No log	6.0	324	0.0649	0.7568	0.8193	0.7868
No log	7.0	378	0.0668	0.7585	0.8113	0.7840
No log	8.0	432	0.0825	0.7094	0.8142	0.7582
No log	9.0	486	0.0767	0.7653	0.8252	0.7941
0.0406	10.0	540	0.0841	0.7621	0.8040	0.7825
0.0406	11.0	594	0.0850	0.7678	0.8032	0.7851
0.0406	12.0	648	0.0880	0.7579	0.7944	0.7757
0.0406	13.0	702	0.0914	0.7619	0.8054	0.7831
0.0406	14.0	756	0.0950	0.7613	0.8003	0.7803
0.0406	15.0	810	0.0960	0.7626	0.8061	0.7838

Framework versions

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