Improved-Arabic-bert-nodropout

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

• Loss: 0.8352
• Accuracy: 0.83

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: 32
• eval_batch_size: 32
• seed: 42
• 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	Accuracy
0.5185	0.55	50	0.3813	0.81
0.3312	1.1	100	0.3438	0.86
0.288	1.65	150	0.3476	0.84
0.2155	2.2	200	0.4957	0.81
0.1593	2.75	250	0.4748	0.81
0.1352	3.3	300	0.5133	0.8
0.0887	3.85	350	0.4520	0.86
0.0497	4.4	400	0.7308	0.82
0.0802	4.95	450	0.6633	0.84
0.0188	5.49	500	0.6727	0.85
0.0384	6.04	550	0.8719	0.85
0.0251	6.59	600	0.7503	0.86
0.0193	7.14	650	0.8141	0.84
0.0196	7.69	700	0.7246	0.86
0.014	8.24	750	0.8134	0.83
0.0155	8.79	800	0.7549	0.88
0.0104	9.34	850	0.7811	0.83
0.01	9.89	900	0.8459	0.83
0.004	10.44	950	0.7777	0.88
0.0007	10.99	1000	0.8015	0.87
0.0021	11.54	1050	0.7967	0.87
0.0009	12.09	1100	0.8019	0.88
0.0111	12.64	1150	0.8450	0.87
0.0003	13.19	1200	0.8463	0.84
0.0022	13.74	1250	0.8715	0.85
0.0039	14.29	1300	0.8378	0.83
0.0023	14.84	1350	0.8352	0.83

Framework versions

• Transformers 4.34.1
• Pytorch 2.1.0+cu118
• Datasets 2.14.7
• Tokenizers 0.14.1