bert-base-uncased_legal_ner_finetuned

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

• Loss: 0.2878
• Law Precision: 0.7983
• Law Recall: 0.8716
• Law F1: 0.8333
• Law Number: 109
• Violated by Precision: 0.7681
• Violated by Recall: 0.7465
• Violated by F1: 0.7571
• Violated by Number: 71
• Violated on Precision: 0.4143
• Violated on Recall: 0.4143
• Violated on F1: 0.4143
• Violated on Number: 70
• Violation Precision: 0.59
• Violation Recall: 0.6941
• Violation F1: 0.6378
• Violation Number: 425
• Overall Precision: 0.6227
• Overall Recall: 0.6993
• Overall F1: 0.6588
• Overall Accuracy: 0.9462

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: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Law Precision	Law Recall	Law F1	Law Number	Violated by Precision	Violated by Recall	Violated by F1	Violated by Number	Violated on Precision	Violated on Recall	Violated on F1	Violated on Number	Violation Precision	Violation Recall	Violation F1	Violation Number	Overall Precision	Overall Recall	Overall F1	Overall Accuracy
No log	1.0	85	0.8260	0.0	0.0	0.0	109	0.0	0.0	0.0	71	0.0	0.0	0.0	70	0.0	0.0	0.0	425	0.0	0.0	0.0	0.7656
No log	2.0	170	0.4451	0.0	0.0	0.0	109	0.0	0.0	0.0	71	0.0	0.0	0.0	70	0.1204	0.1624	0.1383	425	0.1204	0.1022	0.1106	0.8766
No log	3.0	255	0.3153	0.1724	0.0917	0.1198	109	0.0	0.0	0.0	71	0.0	0.0	0.0	70	0.3142	0.36	0.3355	425	0.2991	0.2415	0.2672	0.9067
No log	4.0	340	0.2416	0.6574	0.6514	0.6544	109	0.0	0.0	0.0	71	0.16	0.0571	0.0842	70	0.4496	0.5671	0.5016	425	0.4470	0.4681	0.4573	0.9286
No log	5.0	425	0.2185	0.7768	0.7982	0.7873	109	0.6491	0.5211	0.5781	71	0.3125	0.2857	0.2985	70	0.5019	0.6329	0.5598	425	0.5371	0.6119	0.5720	0.9412
0.5331	6.0	510	0.2399	0.6767	0.8257	0.7438	109	0.6842	0.7324	0.7075	71	0.2841	0.3571	0.3165	70	0.5820	0.7012	0.6361	425	0.5748	0.6889	0.6267	0.9416
0.5331	7.0	595	0.2407	0.7603	0.8440	0.8	109	0.7286	0.7183	0.7234	71	0.4348	0.4286	0.4317	70	0.5752	0.6753	0.6212	425	0.6061	0.6815	0.6416	0.9441
0.5331	8.0	680	0.2610	0.7661	0.8716	0.8155	109	0.6	0.7606	0.6708	71	0.3043	0.4	0.3457	70	0.5948	0.7012	0.6436	425	0.5886	0.7037	0.6410	0.9428
0.5331	9.0	765	0.2790	0.744	0.8532	0.7949	109	0.8667	0.7324	0.7939	71	0.3788	0.3571	0.3676	70	0.5812	0.6824	0.6277	425	0.6133	0.6815	0.6456	0.9461
0.5331	10.0	850	0.2878	0.7983	0.8716	0.8333	109	0.7681	0.7465	0.7571	71	0.4143	0.4143	0.4143	70	0.59	0.6941	0.6378	425	0.6227	0.6993	0.6588	0.9462

Framework versions

• Transformers 4.44.0
• Pytorch 2.4.0
• Datasets 2.21.0
• Tokenizers 0.19.1