kanuni_model

This model is a fine-tuned version of nlpaueb/legal-bert-base-uncased on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.4509
• Threatening Accuracy: 0.75
• Threatening F1: 0.75
• Threatening Precision: 1.0
• Threatening Recall: 0.6
• Conciliatory Accuracy: 0.8125
• Conciliatory F1: 0.8
• Conciliatory Precision: 0.6667
• Conciliatory Recall: 1.0
• Formal Accuracy: 1.0
• Formal F1: 1.0
• Formal Precision: 1.0
• Formal Recall: 1.0
• Persuasive Accuracy: 0.6875
• Persuasive F1: 0.7619
• Persuasive Precision: 0.7273
• Persuasive Recall: 0.8
• Overall Accuracy: 0.8125

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: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 5
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Threatening Accuracy	Threatening F1	Threatening Precision	Threatening Recall	Conciliatory Accuracy	Conciliatory F1	Conciliatory Precision	Conciliatory Recall	Formal Accuracy	Formal F1	Formal Precision	Formal Recall	Persuasive Accuracy	Persuasive F1	Persuasive Precision	Persuasive Recall	Overall Accuracy
0.6671	1.0	27	0.6366	0.4595	0.0909	1.0	0.0476	0.6216	0.2632	0.5	0.1786	0.7297	0.8438	0.7297	1.0	0.6892	0.7723	0.6842	0.8864	0.625
0.6124	2.0	54	0.5715	0.4865	0.1739	1.0	0.0952	0.6351	0.5424	0.5161	0.5714	0.8919	0.9298	0.8833	0.9815	0.7027	0.78	0.6964	0.8864	0.6791
0.531	3.0	81	0.5137	0.7297	0.7222	0.8667	0.6190	0.6892	0.6102	0.5806	0.6429	0.9730	0.9811	1.0	0.9630	0.6892	0.7294	0.7561	0.7045	0.7703
0.4845	4.0	108	0.4720	0.7568	0.7568	0.875	0.6667	0.7703	0.7385	0.6486	0.8571	0.9730	0.9811	1.0	0.9630	0.7027	0.7556	0.7391	0.7727	0.8007
0.4287	5.0	135	0.4524	0.7162	0.7273	0.8	0.6667	0.7432	0.6984	0.6286	0.7857	0.9730	0.9811	1.0	0.9630	0.6892	0.7294	0.7561	0.7045	0.7804

Framework versions

• Transformers 4.48.3
• Pytorch 2.5.1+cu124
• Datasets 3.3.2
• Tokenizers 0.21.0