distilbert_agnews_padding60model

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

• Loss: 0.6521
• Accuracy: 0.9441

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
0.1838	1.0	7500	0.1831	0.9422
0.1417	2.0	15000	0.2010	0.9418
0.1226	3.0	22500	0.2138	0.9476
0.0872	4.0	30000	0.2581	0.9426
0.0614	5.0	37500	0.2946	0.9421
0.0432	6.0	45000	0.3639	0.9420
0.0384	7.0	52500	0.4363	0.9382
0.0297	8.0	60000	0.3892	0.9436
0.0206	9.0	67500	0.4465	0.9408
0.0126	10.0	75000	0.4713	0.9414
0.0146	11.0	82500	0.4982	0.9432
0.0143	12.0	90000	0.5465	0.9392
0.0092	13.0	97500	0.5421	0.9382
0.0055	14.0	105000	0.5625	0.9404
0.0025	15.0	112500	0.6237	0.9416
0.0014	16.0	120000	0.5890	0.9434
0.0065	17.0	127500	0.6210	0.9458
0.003	18.0	135000	0.6399	0.9439
0.0011	19.0	142500	0.6500	0.9458
0.0006	20.0	150000	0.6521	0.9441

Framework versions

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