AraBERT_multi_token_classification_AraEval24_non_dupl_new_labels_set2

This model is a fine-tuned version of aubmindlab/bert-base-arabertv2 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.5971
• Precision: 0.1681
• Recall: 0.1438
• F1: 0.1550
• Accuracy: 0.6956

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: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
1.2013	1.0	776	1.2141	0.0741	0.0072	0.0131	0.7377
0.9773	2.0	1552	1.1423	0.1333	0.0442	0.0664	0.7334
0.8187	3.0	2328	1.1962	0.1562	0.0832	0.1086	0.7191
0.6377	4.0	3104	1.2847	0.1515	0.1125	0.1291	0.6886
0.5292	5.0	3880	1.3504	0.1831	0.1212	0.1459	0.7167
0.4089	6.0	4656	1.4482	0.1697	0.1371	0.1517	0.7046
0.3487	7.0	5432	1.4719	0.1591	0.1366	0.1470	0.7070
0.285	8.0	6208	1.5242	0.1685	0.1510	0.1593	0.6995
0.2643	9.0	6984	1.5617	0.1720	0.1397	0.1542	0.7013
0.224	10.0	7760	1.5971	0.1681	0.1438	0.1550	0.6956

Framework versions

• Transformers 4.30.2
• Pytorch 2.3.1+cu121
• Datasets 2.20.0
• Tokenizers 0.13.3