N_bert_agnews_padding40model

This model is a fine-tuned version of bert-base-uncased on the ag_news dataset. It achieves the following results on the evaluation set:

• Loss: 0.5661
• Accuracy: 0.9474

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: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.1785	1.0	7500	0.1884	0.9421
0.1379	2.0	15000	0.1990	0.9478
0.1127	3.0	22500	0.2389	0.9408
0.0846	4.0	30000	0.2528	0.9492
0.0581	5.0	37500	0.3041	0.9436
0.0456	6.0	45000	0.3415	0.9468
0.0411	7.0	52500	0.4081	0.9430
0.0239	8.0	60000	0.4415	0.9433
0.0202	9.0	67500	0.4380	0.9404
0.0126	10.0	75000	0.4637	0.9425
0.0175	11.0	82500	0.4485	0.9455
0.0126	12.0	90000	0.4761	0.9449
0.0046	13.0	97500	0.5009	0.9455
0.0038	14.0	105000	0.4784	0.9482
0.0035	15.0	112500	0.5282	0.9451
0.0046	16.0	120000	0.5256	0.9464
0.0026	17.0	127500	0.5081	0.9501
0.0008	18.0	135000	0.5543	0.9467
0.0002	19.0	142500	0.5448	0.9488
0.0016	20.0	150000	0.5661	0.9474

Framework versions

• Transformers 4.33.2
• Pytorch 2.0.1+cu117
• Datasets 2.14.5
• Tokenizers 0.13.3