CS221-roberta-base-finetuned-augmentation

This model is a fine-tuned version of FacebookAI/roberta-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.1561
• F1: 0.9312
• Roc Auc: 0.9459
• Accuracy: 0.8638

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: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 100
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	F1	Roc Auc	Accuracy
0.3933	1.0	180	0.3696	0.6352	0.7378	0.4288
0.2905	2.0	360	0.2958	0.7524	0.8079	0.5587
0.2139	3.0	540	0.2298	0.8538	0.8845	0.6769
0.1505	4.0	720	0.1996	0.8757	0.9030	0.7429
0.1218	5.0	900	0.1745	0.9027	0.9200	0.7880
0.0737	6.0	1080	0.1556	0.9199	0.9356	0.8290
0.0604	7.0	1260	0.1541	0.9241	0.9366	0.8485
0.0353	8.0	1440	0.1538	0.9294	0.9468	0.8652
0.024	9.0	1620	0.1561	0.9312	0.9459	0.8638
0.0235	10.0	1800	0.1638	0.9283	0.9459	0.8624
0.0159	11.0	1980	0.1615	0.9310	0.9467	0.8694
0.0158	12.0	2160	0.1625	0.9307	0.9461	0.8687

Framework versions

• Transformers 4.47.0
• Pytorch 2.5.1+cu121
• Datasets 3.2.0
• Tokenizers 0.21.0