UIT-NO-PREdeberta-v3-large-finetuned

This model is a fine-tuned version of microsoft/deberta-v3-large on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.7092
• F1: 0.7611
• Roc Auc: 0.8161
• Accuracy: 0.5162

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: 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
0.5182	1.0	139	0.4623	0.4534	0.6457	0.3087
0.3803	2.0	278	0.3662	0.7072	0.7767	0.4061
0.2764	3.0	417	0.3630	0.7128	0.7816	0.4729
0.1508	4.0	556	0.3983	0.7345	0.7944	0.4585
0.124	5.0	695	0.4009	0.7591	0.8210	0.5
0.0748	6.0	834	0.4494	0.7575	0.8153	0.4964
0.0458	7.0	973	0.5271	0.7462	0.8013	0.5018
0.0356	8.0	1112	0.5749	0.7515	0.8060	0.5018
0.0158	9.0	1251	0.6486	0.7290	0.7894	0.4765
0.013	10.0	1390	0.6190	0.7552	0.8133	0.5054
0.0052	11.0	1529	0.6594	0.7571	0.8121	0.5162
0.0024	12.0	1668	0.6776	0.7533	0.8128	0.5018
0.0016	13.0	1807	0.6816	0.7590	0.8154	0.5108
0.0014	14.0	1946	0.6997	0.7597	0.8152	0.5144
0.0017	15.0	2085	0.7081	0.7572	0.8121	0.5162
0.0013	16.0	2224	0.7061	0.7607	0.8161	0.5126
0.0056	17.0	2363	0.7092	0.7611	0.8161	0.5162
0.0013	18.0	2502	0.7142	0.7609	0.8155	0.5162
0.0013	19.0	2641	0.7157	0.7609	0.8155	0.5162
0.0012	20.0	2780	0.7158	0.7609	0.8155	0.5162

Framework versions

• Transformers 4.48.1
• Pytorch 2.4.0
• Datasets 3.0.1
• Tokenizers 0.21.0