Arabic-MARBERT-dialect-Identification-City-Model

This model is a fine-tuned version of Ammar-alhaj-ali/arabic-MARBERT-dialect-identification-city on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 2.8911
• F1-score: 0.5786

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: 5e-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: 20

Training results

Training Loss	Epoch	Step	Validation Loss	F1-score
No log	1.0	442	1.5955	0.4477
1.8911	2.0	884	1.6004	0.5083
1.3056	3.0	1326	1.5384	0.5537
0.989	4.0	1768	1.7560	0.5450
0.7526	5.0	2210	1.8035	0.5572
0.5765	6.0	2652	1.8953	0.5764
0.4799	7.0	3094	2.3015	0.5486
0.3629	8.0	3536	2.7364	0.5368
0.3629	9.0	3978	2.8416	0.5442
0.277	10.0	4420	2.8911	0.5786
0.2196	11.0	4862	3.2772	0.5304
0.1656	12.0	5304	3.1206	0.5394
0.1555	13.0	5746	3.4600	0.5447
0.1315	14.0	6188	3.4378	0.5445
0.1039	15.0	6630	3.5247	0.5409
0.0796	16.0	7072	3.6359	0.5419
0.0683	17.0	7514	3.6723	0.5339
0.0683	18.0	7956	3.6808	0.5435
0.0567	19.0	8398	3.7677	0.5307
0.0382	20.0	8840	3.8121	0.5306

Framework versions

• Transformers 4.40.0
• Pytorch 2.2.1+cu121
• Datasets 2.19.0
• Tokenizers 0.19.1