Modelpublic

MM2157/AraBERT_token_classification_AraEval24_multi_n_dupl_new_labels_aug_back_trans

sourceHugging Faceupdated 2y agoView on Hugging Face

0likes6downloads

Model Card

AraBERTtokenclassificationAraEval24multinduplnewlabelsaugback_trans

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: 1.7261
—Precision: 0.1857
—Recall: 0.1294
—F1: 0.1525
—Accuracy: 0.7083

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
—trainbatchsize: 8
—evalbatchsize: 8
—seed: 42
—optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
—lrschedulertype: linear
—num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
0.9613	1.0	1505	1.2156	0.1877	0.0649	0.0965	0.7232
0.7791	2.0	3010	1.2402	0.2023	0.0817	0.1164	0.7202
0.5796	3.0	4515	1.3461	0.1791	0.0817	0.1122	0.7010
0.4597	4.0	6020	1.4365	0.1676	0.0903	0.1174	0.7037
0.3678	5.0	7525	1.5251	0.1756	0.1126	0.1372	0.7015
0.3085	6.0	9030	1.5398	0.1886	0.1563	0.1709	0.6866
0.2707	7.0	10535	1.6344	0.1831	0.1553	0.1680	0.7030
0.2427	8.0	12040	1.6892	0.1743	0.1177	0.1405	0.7088
0.2116	9.0	13545	1.7166	0.1859	0.1218	0.1471	0.7095
0.1977	10.0	15050	1.7261	0.1857	0.1294	0.1525	0.7083

Framework versions

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