Model4_withclasess-arabertv2_base_T2_WS_A100v2_F1__BL

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

• Loss: 0.0770
• F1-micro: 0.8282
• Roc Auc: 0.9072
• Accuracy: 0.7912

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: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	F1-micro	Roc Auc	Accuracy
0.1028	1.0	507	0.0644	0.7904	0.8639	0.7297
0.0476	2.0	1014	0.0556	0.8143	0.8828	0.7668
0.0308	3.0	1521	0.0570	0.8200	0.8929	0.7758
0.0206	4.0	2028	0.0624	0.8179	0.8979	0.7828
0.0134	5.0	2535	0.0696	0.8183	0.9016	0.7856
0.0097	6.0	3042	0.0743	0.8226	0.9052	0.7898
0.0077	7.0	3549	0.0779	0.8166	0.9039	0.7793
0.0054	8.0	4056	0.0809	0.8249	0.9063	0.7905
0.0045	9.0	4563	0.0770	0.8282	0.9072	0.7912
0.0036	10.0	5070	0.0812	0.8228	0.9049	0.7849
0.003	11.0	5577	0.0874	0.8250	0.9072	0.7919
0.0025	12.0	6084	0.0886	0.8258	0.9067	0.7863

Framework versions

• Transformers 4.48.2
• Pytorch 2.5.1+cu124
• Tokenizers 0.21.0