N_bert_agnews_padding0model

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: 0.5653
• Accuracy: 0.9476

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
0.1773	1.0	7500	0.1841	0.9433
0.1354	2.0	15000	0.2061	0.9463
0.1138	3.0	22500	0.2455	0.9428
0.0852	4.0	30000	0.2881	0.9429
0.0627	5.0	37500	0.3271	0.9433
0.0436	6.0	45000	0.3524	0.9441
0.034	7.0	52500	0.3977	0.9424
0.0251	8.0	60000	0.4291	0.9441
0.0205	9.0	67500	0.4399	0.9420
0.0167	10.0	75000	0.4574	0.9429
0.0218	11.0	82500	0.4979	0.9429
0.0119	12.0	90000	0.5000	0.9438
0.0112	13.0	97500	0.4856	0.9454
0.0054	14.0	105000	0.5294	0.9457
0.0039	15.0	112500	0.5418	0.9459
0.0024	16.0	120000	0.5065	0.9468
0.0011	17.0	127500	0.5511	0.9458
0.0013	18.0	135000	0.5411	0.9471
0.0002	19.0	142500	0.5555	0.9472
0.0005	20.0	150000	0.5653	0.9476

Framework versions

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