resultmbert2mbert

This model is a fine-tuned version of malmarjeh/mbert2mbert-arabic-text-summarization on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.8701

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: 32
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
2.551	0.4263	500	1.0592
1.1939	0.8525	1000	0.9787
1.0979	1.2788	1500	0.9425
1.0436	1.7050	2000	0.9134
1.0132	2.1313	2500	0.9038
0.9645	2.5575	3000	0.8905
0.9608	2.9838	3500	0.8857
0.9526	3.4101	4000	0.8931
0.96	3.8363	4500	0.8838
0.9254	4.2626	5000	0.8804
0.9023	4.6888	5500	0.8724
0.884	5.1151	6000	0.8754
0.8496	5.5413	6500	0.8656
0.85	5.9676	7000	0.8653
0.8076	6.3939	7500	0.8668
0.8119	6.8201	8000	0.8655
0.7953	7.2464	8500	0.8676
0.7719	7.6726	9000	0.8656
0.7657	8.0989	9500	0.8710
0.7446	8.5251	10000	0.8694
0.7524	8.9514	10500	0.8658
0.729	9.3777	11000	0.8699
0.7338	9.8039	11500	0.8701

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.0+cu121
• Datasets 3.0.0
• Tokenizers 0.19.1