trainer_output

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

• Loss: 2.0023
• Accuracy: 0.8598
• F1: 0.8668
• Precision: 0.8770
• Recall: 0.8598

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: 2e-05
• train_batch_size: 16
• eval_batch_size: 32
• seed: 42
• 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
• num_epochs: 3
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall
No log	0.2439	20	1.6724	0.8537	0.8597	0.8721	0.8537
No log	0.4878	40	2.1847	0.8598	0.8618	0.8640	0.8598
0.3126	0.7317	60	2.0168	0.8598	0.8630	0.8673	0.8598
0.3126	0.9756	80	2.4119	0.8780	0.8804	0.8904	0.8780
0.2521	1.2195	100	2.2020	0.8902	0.8884	0.8899	0.8902
0.2521	1.4634	120	2.2252	0.8902	0.8893	0.8908	0.8902
0.2521	1.7073	140	1.9534	0.8476	0.8579	0.8738	0.8476
0.2102	1.9512	160	2.0566	0.8963	0.8952	0.8948	0.8963
0.2102	2.1951	180	2.1647	0.8659	0.8714	0.8799	0.8659
0.0475	2.4390	200	2.2178	0.8659	0.8713	0.8795	0.8659
0.0475	2.6829	220	2.2616	0.8659	0.8713	0.8795	0.8659
0.0475	2.9268	240	2.2667	0.8659	0.8713	0.8795	0.8659

Framework versions

• Transformers 4.56.2
• Pytorch 2.8.0+cu126
• Datasets 4.0.0
• Tokenizers 0.22.1