klue-roberta-large-klue-crime-2-ner

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

• Loss: 0.5080
• Precision: 0.0727
• Recall: 0.0449
• F1: 0.0556
• Accuracy: 0.8794

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-06
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	1.0	7	1.0363	0.0	0.0	0.0	0.8425
No log	2.0	14	0.8076	0.0	0.0	0.0	0.8501
No log	3.0	21	0.6843	0.0	0.0	0.0	0.8554
No log	4.0	28	0.6184	0.0	0.0	0.0	0.8595
No log	5.0	35	0.5803	0.0	0.0	0.0	0.8618
No log	6.0	42	0.5546	0.0	0.0	0.0	0.8706
No log	7.0	49	0.5338	0.0189	0.0112	0.0141	0.8741
No log	8.0	56	0.5195	0.0370	0.0225	0.0280	0.8753
No log	9.0	63	0.5111	0.0545	0.0337	0.0417	0.8770
No log	10.0	70	0.5080	0.0727	0.0449	0.0556	0.8794

Framework versions

• Transformers 4.30.2
• Pytorch 2.0.1+cpu
• Datasets 2.12.0
• Tokenizers 0.11.0