MArabizi_Checkpoints_v444_v6

This model is a fine-tuned version of Mohamedd123321/Arabizi_Checkpoints_v444 on an unknown dataset. It achieves the following results on the evaluation set:

• Accuracy: 0.7630
• F1: 0.7509
• Loss: 0.7956
• Precision: 0.7577
• Recall: 0.7630

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: 1e-05
• train_batch_size: 8
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.06
• num_epochs: 6

Training results

Training Loss	Epoch	Step	Accuracy	F1	Validation Loss	Precision	Recall
0.4959	0.8258	2000	0.7583	0.7376	0.5365	0.7605	0.7583
0.4054	1.6516	4000	0.7605	0.7411	0.5825	0.7519	0.7605
0.3405	2.4774	6000	0.7677	0.7530	0.6143	0.7644	0.7677
0.2731	3.3032	8000	0.7599	0.7418	0.6971	0.7568	0.7599
0.2343	4.1290	10000	0.7576	0.7398	0.7517	0.7528	0.7576
0.2341	4.9563	12000	0.7620	0.7488	0.7525	0.7570	0.7620
0.1905	5.7821	14000	0.7630	0.7509	0.7956	0.7577	0.7630

Framework versions

• Transformers 4.40.2
• Pytorch 2.8.0+cu129
• Datasets 5.0.1
• Tokenizers 0.19.1