Improved-Arabert-twitter-sentiment-No-dropout

This model is a fine-tuned version of aubmindlab/bert-base-arabertv2 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.5434
• Accuracy: 0.9

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
0.5764	0.55	50	0.4925	0.79
0.3357	1.1	100	0.3094	0.88
0.2183	1.65	150	0.2971	0.87
0.2042	2.2	200	0.3013	0.89
0.1258	2.75	250	0.3038	0.9
0.1359	3.3	300	0.3114	0.89
0.0893	3.85	350	0.3108	0.91
0.0816	4.4	400	0.3569	0.9
0.071	4.95	450	0.3574	0.9
0.1043	5.49	500	0.4332	0.89
0.0407	6.04	550	0.4232	0.9
0.0378	6.59	600	0.4273	0.91
0.0341	7.14	650	0.4671	0.91
0.0226	7.69	700	0.5174	0.9
0.0215	8.24	750	0.4786	0.89
0.0329	8.79	800	0.4853	0.9
0.021	9.34	850	0.5430	0.9
0.0219	9.89	900	0.5510	0.89
0.0154	10.44	950	0.5518	0.9
0.0119	10.99	1000	0.5473	0.9
0.0108	11.54	1050	0.5285	0.9
0.0138	12.09	1100	0.5239	0.91
0.0133	12.64	1150	0.5584	0.89
0.0121	13.19	1200	0.5334	0.9
0.0063	13.74	1250	0.5325	0.91
0.0061	14.29	1300	0.5429	0.9
0.0105	14.84	1350	0.5434	0.9

Framework versions

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