roberta-base_refusal

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

• Loss: 0.0104
• Accuracy: 0.9981

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: 5e-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
• lr_scheduler_warmup_steps: 500
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.0002	0.2776	500	0.0159	0.9975
0.0011	0.5552	1000	0.0372	0.9931
0.0376	0.8329	1500	0.0214	0.9963
0.038	1.1105	2000	0.0147	0.9981
0.0465	1.3881	2500	0.0157	0.9975
0.0495	1.6657	3000	0.0096	0.9988
0.0011	1.9434	3500	0.0136	0.9981
0.0004	2.2210	4000	0.0153	0.9981
0.0007	2.4986	4500	0.0123	0.9981
0.0004	2.7762	5000	0.0362	0.9956
0.044	3.0539	5500	0.0176	0.9975
0.0015	3.3315	6000	0.0142	0.9975
0.0004	3.6091	6500	0.0113	0.9981
0.0004	3.8867	7000	0.0227	0.9956
0.0436	4.1644	7500	0.0129	0.9969
0.0085	4.4420	8000	0.0102	0.9988
0.0003	4.7196	8500	0.0102	0.9988
0.0002	4.9972	9000	0.0104	0.9981

Framework versions

• Transformers 4.42.3
• Pytorch 2.1.2
• Datasets 2.20.0
• Tokenizers 0.19.1