N_roberta_agnews_padding100model

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

• Loss: 0.5447
• Accuracy: 0.95

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
0.1985	1.0	7500	0.2020	0.9422
0.1646	2.0	15000	0.2020	0.9467
0.1491	3.0	22500	0.2176	0.9462
0.1251	4.0	30000	0.2385	0.9486
0.1071	5.0	37500	0.2422	0.9479
0.0842	6.0	45000	0.2795	0.9470
0.0728	7.0	52500	0.3227	0.9429
0.0558	8.0	60000	0.3396	0.9462
0.0493	9.0	67500	0.3946	0.9454
0.0406	10.0	75000	0.3891	0.9471
0.026	11.0	82500	0.4082	0.9492
0.0211	12.0	90000	0.4271	0.9454
0.0176	13.0	97500	0.4244	0.9468
0.0114	14.0	105000	0.4723	0.9467
0.0116	15.0	112500	0.4950	0.9459
0.0097	16.0	120000	0.4863	0.9501
0.0098	17.0	127500	0.4869	0.9496
0.0046	18.0	135000	0.4984	0.9516
0.0008	19.0	142500	0.5340	0.9491
0.0011	20.0	150000	0.5447	0.95

Framework versions

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