afro-xlmr-base-hau-finetuned-augmentation-LUNAR

This model is a fine-tuned version of Davlan/afro-xlmr-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.3766
• F1: 0.7024
• Roc Auc: 0.8104
• Accuracy: 0.5513

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
0.4359	1.0	144	0.4259	0.1730	0.5612	0.1965
0.3509	2.0	288	0.3506	0.4649	0.6710	0.3670
0.3062	3.0	432	0.3230	0.5924	0.7446	0.4452
0.2507	4.0	576	0.3000	0.6340	0.7629	0.4748
0.2156	5.0	720	0.2989	0.6716	0.7870	0.5235
0.1632	6.0	864	0.3159	0.6645	0.7827	0.5009
0.1665	7.0	1008	0.3168	0.6817	0.7973	0.5217
0.1289	8.0	1152	0.3148	0.6807	0.7958	0.5357
0.1166	9.0	1296	0.3261	0.6850	0.7946	0.5217
0.0927	10.0	1440	0.3268	0.6828	0.7910	0.5496
0.0693	11.0	1584	0.3387	0.6982	0.8028	0.5496
0.0571	12.0	1728	0.3544	0.6938	0.8050	0.5374
0.0568	13.0	1872	0.3439	0.6959	0.8037	0.5565
0.0467	14.0	2016	0.3673	0.6940	0.8054	0.5391
0.043	15.0	2160	0.3642	0.7015	0.8066	0.5548
0.0356	16.0	2304	0.3685	0.6998	0.8069	0.5530
0.0395	17.0	2448	0.3766	0.7024	0.8104	0.5513
0.0353	18.0	2592	0.3766	0.7005	0.8088	0.5478
0.0352	19.0	2736	0.3753	0.7008	0.8092	0.5478
0.0327	20.0	2880	0.3745	0.6985	0.8074	0.5461

Framework versions

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