roberta-large-finetuned-augmentation

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

• Loss: 0.3233
• F1: 0.8669
• Roc Auc: 0.9017
• Accuracy: 0.7292

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: cosine
• lr_scheduler_warmup_steps: 100
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	F1	Roc Auc	Accuracy
0.2899	1.0	421	0.2622	0.8173	0.8503	0.6039
0.2244	2.0	842	0.2303	0.8472	0.8884	0.6752
0.1662	3.0	1263	0.2332	0.8534	0.8863	0.7078
0.12	4.0	1684	0.2640	0.8517	0.8829	0.6971
0.1192	5.0	2105	0.2712	0.8516	0.8866	0.6954
0.0805	6.0	2526	0.2858	0.8614	0.8967	0.7162
0.0918	7.0	2947	0.2726	0.8625	0.8932	0.7233
0.0814	8.0	3368	0.3005	0.8639	0.8964	0.7090
0.0732	9.0	3789	0.3067	0.8555	0.8981	0.7055
0.0511	10.0	4210	0.2953	0.8645	0.9001	0.7203
0.0516	11.0	4631	0.3022	0.8623	0.8942	0.7167
0.0574	12.0	5052	0.3061	0.8658	0.9024	0.7251
0.0549	13.0	5473	0.3151	0.8625	0.8972	0.7167
0.0387	14.0	5894	0.3201	0.8669	0.9016	0.7274
0.0468	15.0	6315	0.3178	0.8669	0.9028	0.7221
0.0302	16.0	6736	0.3225	0.8658	0.8985	0.7251
0.0332	17.0	7157	0.3233	0.8669	0.9017	0.7292
0.0272	18.0	7578	0.3261	0.8652	0.9010	0.7245
0.0256	19.0	7999	0.3275	0.8656	0.9015	0.7245
0.0273	20.0	8420	0.3275	0.8654	0.9009	0.7251

Framework versions

• Transformers 4.45.1
• Pytorch 2.4.0
• Datasets 3.0.1
• Tokenizers 0.20.0