Type_of_relation

This model is a fine-tuned version of aubmindlab/bert-base-arabertv02 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.3733
• Macro F1: 0.7820
• Precision: 0.7773
• Recall: 0.7917
• Kappa: 0.6875
• Accuracy: 0.7917

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: 128
• seed: 25
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 15

Training results

Training Loss	Epoch	Step	Validation Loss	Macro F1	Precision	Recall	Kappa	Accuracy
1.1229	1.0	697	0.8030	0.7500	0.7330	0.7852	0.6631	0.7852
0.8056	2.0	1395	0.7428	0.7698	0.7571	0.7974	0.6888	0.7974
0.64	3.0	2092	0.7285	0.7731	0.7653	0.7958	0.6867	0.7958
0.5323	4.0	2790	0.7310	0.7848	0.7758	0.8019	0.7010	0.8019
0.4717	5.0	3487	0.8139	0.7777	0.7736	0.7924	0.6849	0.7924
0.312	6.0	4185	0.8625	0.7835	0.7761	0.7974	0.6950	0.7974
0.2707	7.0	4882	0.9528	0.7824	0.7804	0.7910	0.6869	0.7910
0.1907	8.0	5580	1.0535	0.7814	0.7749	0.7962	0.6902	0.7962
0.1494	9.0	6277	1.1044	0.7791	0.7761	0.7863	0.6825	0.7863
0.1408	10.0	6975	1.1593	0.7818	0.7790	0.7879	0.6845	0.7879
0.0949	11.0	7672	1.2428	0.7846	0.7791	0.7954	0.6920	0.7954
0.0815	12.0	8370	1.2998	0.7834	0.7770	0.7963	0.6926	0.7963
0.0657	13.0	9067	1.3431	0.7827	0.7784	0.7929	0.6889	0.7929
0.0509	14.0	9765	1.3687	0.7813	0.7773	0.7910	0.6863	0.7910
0.0488	14.99	10455	1.3733	0.7820	0.7773	0.7917	0.6875	0.7917

Framework versions

• Transformers 4.36.0
• Pytorch 2.1.0+cu118
• Tokenizers 0.15.0