distilbert-base-multilingual-cased_dna_ro_m12

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

• Loss: 0.0541
• Accuracy: 0.982
• Accuracy Refusal: 0.997
• Precision: 0.976
• Precision Refusal: 0.975
• Recall: 0.98
• Recall Refusal: 0.99
• F1: 0.978
• F1 Refusal: 0.982

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: 50
• eval_batch_size: 100
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Accuracy Refusal	Precision	Precision Refusal	Recall	Recall Refusal	F1	F1 Refusal
0.3288	1.0	1465	0.2399	0.917	0.989	0.888	0.941	0.9	0.916	0.893	0.928
0.1885	2.0	2930	0.1140	0.96	0.993	0.95	0.937	0.953	0.984	0.951	0.959
0.1253	3.0	4395	0.0541	0.982	0.997	0.976	0.975	0.98	0.99	0.978	0.982

Framework versions

• Transformers 4.43.1
• Pytorch 2.3.0+cu118
• Datasets 2.16.1
• Tokenizers 0.19.1