ar-poem-classification

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: 1.1427
• Macro F1: 0.6954
• Accuracy: 0.6944

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: 5
• total_train_batch_size: 80
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 12

Training results

Training Loss	Epoch	Step	Validation Loss	Macro F1	Accuracy
No log	1.0	250	1.0896	0.5319	0.5334
1.114	2.0	500	0.9989	0.5864	0.5826
1.114	3.0	750	0.9993	0.5942	0.5976
0.8219	4.0	1000	0.9949	0.6042	0.609
0.8219	5.0	1250	0.9813	0.6337	0.6366
0.563	6.0	1500	0.9666	0.6657	0.6654
0.563	7.0	1750	1.0253	0.6686	0.6668
0.3763	8.0	2000	1.0150	0.6951	0.6936
0.3763	9.0	2250	1.0619	0.6872	0.6872
0.2525	10.0	2500	1.1035	0.6929	0.6922
0.2525	11.0	2750	1.1352	0.6952	0.6944
0.184	12.0	3000	1.1427	0.6954	0.6944

Framework versions

• Transformers 4.38.2
• Pytorch 2.1.0+cu121
• Tokenizers 0.15.2