xlm-roberta-base-amh-finetuned-augmentation-LUNAR

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

• Loss: 0.3820
• F1: 0.6576
• Roc Auc: 0.7940
• Accuracy: 0.5110

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: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 100
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	F1	Roc Auc	Accuracy
0.4179	1.0	238	0.4173	0.0	0.5	0.1735
0.3969	2.0	476	0.3760	0.2222	0.5646	0.2808
0.3427	3.0	714	0.3236	0.4125	0.6565	0.3922
0.2814	4.0	952	0.2987	0.5339	0.7071	0.4448
0.2655	5.0	1190	0.2953	0.5704	0.7415	0.4658
0.2185	6.0	1428	0.3105	0.5699	0.7519	0.4448
0.2206	7.0	1666	0.3094	0.5764	0.7509	0.4669
0.161	8.0	1904	0.3088	0.5871	0.7524	0.4879
0.1473	9.0	2142	0.3198	0.6278	0.7683	0.4921
0.1237	10.0	2380	0.3405	0.6264	0.7680	0.4774
0.1032	11.0	2618	0.3341	0.6362	0.7720	0.5079
0.0857	12.0	2856	0.3452	0.6521	0.7771	0.5058
0.0695	13.0	3094	0.3604	0.6552	0.7872	0.5058
0.0626	14.0	3332	0.3686	0.6472	0.7815	0.5089
0.0481	15.0	3570	0.3666	0.6477	0.7763	0.5121
0.0516	16.0	3808	0.3820	0.6576	0.7940	0.5110
0.0469	17.0	4046	0.3752	0.6493	0.7846	0.5121
0.0473	18.0	4284	0.3817	0.6448	0.7821	0.5100
0.0405	19.0	4522	0.3830	0.6529	0.7871	0.5110
0.0474	20.0	4760	0.3827	0.6519	0.7874	0.5100

Framework versions

• Transformers 4.45.1
• Pytorch 2.4.0
• Datasets 3.0.1
• Tokenizers 0.20.0