N_distilbert_agnews_padding20model

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: 0.6409
• Accuracy: 0.9447

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
0.1753	1.0	7500	0.1871	0.9414
0.1361	2.0	15000	0.2002	0.9443
0.1127	3.0	22500	0.2222	0.9439
0.0824	4.0	30000	0.2883	0.9411
0.0503	5.0	37500	0.3229	0.9405
0.0379	6.0	45000	0.3573	0.9426
0.0336	7.0	52500	0.3657	0.9421
0.0287	8.0	60000	0.4377	0.9433
0.0174	9.0	67500	0.4701	0.9421
0.0177	10.0	75000	0.4664	0.9436
0.0176	11.0	82500	0.4614	0.9461
0.0106	12.0	90000	0.5123	0.9436
0.0091	13.0	97500	0.5772	0.9428
0.0043	14.0	105000	0.5659	0.9436
0.0048	15.0	112500	0.5934	0.9442
0.0037	16.0	120000	0.5910	0.945
0.0039	17.0	127500	0.6058	0.9461
0.0022	18.0	135000	0.6002	0.9468
0.0025	19.0	142500	0.6316	0.9461
0.0018	20.0	150000	0.6409	0.9447

Framework versions

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