N_bert_agnews_padding100model

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: 0.5818
• Accuracy: 0.945

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
0.1815	1.0	7500	0.1891	0.9421
0.1362	2.0	15000	0.2013	0.9446
0.1152	3.0	22500	0.2381	0.9443
0.0809	4.0	30000	0.2646	0.9453
0.0598	5.0	37500	0.3089	0.9425
0.0405	6.0	45000	0.3708	0.9391
0.0387	7.0	52500	0.3904	0.9418
0.0212	8.0	60000	0.4448	0.9432
0.0225	9.0	67500	0.4465	0.9429
0.0145	10.0	75000	0.4374	0.9445
0.017	11.0	82500	0.4895	0.9438
0.0091	12.0	90000	0.4848	0.9443
0.0128	13.0	97500	0.4764	0.9455
0.0044	14.0	105000	0.5263	0.9449
0.0018	15.0	112500	0.5252	0.9447
0.0017	16.0	120000	0.5324	0.9468
0.001	17.0	127500	0.5503	0.9457
0.0006	18.0	135000	0.5748	0.9458
0.0002	19.0	142500	0.5715	0.9459
0.0015	20.0	150000	0.5818	0.945

Framework versions

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