Metrics

• loss: 1.0628
• accuracy: 0.8708
• precision: 0.8661
• recall: 0.8708
• precision_macro: 0.8180
• recall_macro: 0.6890
• macro_fpr: 0.0681
• weighted_fpr: 0.0471
• weighted_specificity: 0.8788
• macro_specificity: 0.9374
• weighted_sensitivity: 0.8708
• macro_sensitivity: 0.6890
• f1_micro: 0.8708
• f1_macro: 0.7165
• f1_weighted: 0.8586
• runtime: 13.9241
• samples_per_second: 32.2460
• steps_per_second: 4.0940

case-analysis-legal-bert-base-uncased

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: 1.3822
• Accuracy: 0.8263
• Precision: 0.8205
• Recall: 0.8263
• Precision Macro: 0.6455
• Recall Macro: 0.6413
• Macro Fpr: 0.0910
• Weighted Fpr: 0.0732
• Weighted Specificity: 0.8622
• Macro Specificity: 0.9177
• Weighted Sensitivity: 0.8085
• Macro Sensitivity: 0.6413
• F1 Micro: 0.8085
• F1 Macro: 0.6429
• F1 Weighted: 0.8061

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: 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: 30
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	Precision Macro	Recall Macro	Macro Fpr	Weighted Fpr	Weighted Specificity	Macro Specificity	Weighted Sensitivity	Macro Sensitivity	F1 Micro	F1 Macro	F1 Weighted
No log	1.0	224	1.0114	0.6570	0.6179	0.6570	0.4664	0.4075	0.1948	0.1482	0.6727	0.8324	0.6570	0.4075	0.6570	0.4081	0.6166
No log	2.0	448	0.7650	0.7751	0.7425	0.7751	0.5566	0.5806	0.1094	0.0882	0.8406	0.9039	0.7751	0.5806	0.7751	0.5659	0.7564
0.7677	3.0	672	0.7342	0.7817	0.7695	0.7817	0.5674	0.5967	0.1041	0.0851	0.8515	0.9083	0.7817	0.5967	0.7817	0.5707	0.7678
0.7677	4.0	896	0.7968	0.8174	0.7766	0.8174	0.6036	0.5893	0.0965	0.0693	0.8368	0.9136	0.8174	0.5893	0.8174	0.5904	0.7921
0.482	5.0	1120	0.8171	0.8085	0.7853	0.8085	0.6366	0.6038	0.0990	0.0732	0.8346	0.9108	0.8085	0.6038	0.8085	0.6141	0.7940
0.482	6.0	1344	0.8910	0.8241	0.8028	0.8241	0.6660	0.6359	0.0875	0.0664	0.8606	0.9212	0.8241	0.6359	0.8241	0.6315	0.8084
0.2993	7.0	1568	1.0094	0.8040	0.8109	0.8040	0.6742	0.6774	0.0891	0.0751	0.8829	0.9217	0.8040	0.6774	0.8040	0.6742	0.8067
0.2993	8.0	1792	1.1504	0.8107	0.7968	0.8107	0.6228	0.6330	0.0897	0.0722	0.8708	0.9204	0.8107	0.6330	0.8107	0.6229	0.8016
0.1367	9.0	2016	1.2533	0.8062	0.8059	0.8062	0.6537	0.6225	0.0948	0.0742	0.8595	0.9164	0.8062	0.6225	0.8062	0.6360	0.8045
0.1367	10.0	2240	1.2516	0.8174	0.8107	0.8174	0.6621	0.6499	0.0873	0.0693	0.8701	0.9219	0.8174	0.6499	0.8174	0.6554	0.8137
0.1367	11.0	2464	1.3822	0.8263	0.8205	0.8263	0.7085	0.6696	0.0833	0.0655	0.8711	0.9243	0.8263	0.6696	0.8263	0.6764	0.8195
0.055	12.0	2688	1.4574	0.8018	0.8127	0.8018	0.6369	0.6443	0.0883	0.0761	0.8844	0.9216	0.8018	0.6443	0.8018	0.6399	0.8068
0.055	13.0	2912	1.6634	0.7884	0.7810	0.7884	0.6090	0.6042	0.1002	0.0821	0.8619	0.9126	0.7884	0.6042	0.7884	0.6042	0.7831
0.0431	14.0	3136	1.5085	0.8285	0.8077	0.8285	0.6476	0.6367	0.0850	0.0645	0.8633	0.9229	0.8285	0.6367	0.8285	0.6382	0.8166
0.0431	15.0	3360	1.6411	0.8107	0.7936	0.8107	0.6243	0.6262	0.0914	0.0722	0.8626	0.9183	0.8107	0.6262	0.8107	0.6229	0.8014
0.0135	16.0	3584	1.7483	0.8062	0.7925	0.8062	0.6201	0.6271	0.0923	0.0742	0.8647	0.9177	0.8062	0.6271	0.8062	0.6221	0.7988
0.0135	17.0	3808	1.7233	0.7973	0.7897	0.7973	0.6148	0.6263	0.0942	0.0781	0.8682	0.9164	0.7973	0.6263	0.7973	0.6201	0.7933
0.0066	18.0	4032	1.6457	0.8241	0.8042	0.8241	0.6515	0.6388	0.0879	0.0664	0.8522	0.9191	0.8241	0.6388	0.8241	0.6391	0.8115
0.0066	19.0	4256	1.6614	0.8174	0.7976	0.8174	0.6324	0.6420	0.0865	0.0693	0.8703	0.9219	0.8174	0.6420	0.8174	0.6318	0.8061
0.0066	20.0	4480	1.6997	0.8129	0.8023	0.8129	0.6435	0.6576	0.0860	0.0712	0.8759	0.9222	0.8129	0.6576	0.8129	0.6462	0.8061
0.0067	21.0	4704	1.6540	0.8218	0.8000	0.8218	0.6473	0.6380	0.0880	0.0674	0.8560	0.9195	0.8218	0.6380	0.8218	0.6356	0.8088
0.0067	22.0	4928	1.7329	0.8085	0.7945	0.8085	0.6313	0.6267	0.0930	0.0732	0.8548	0.9158	0.8085	0.6267	0.8085	0.6282	0.8011
0.0028	23.0	5152	1.7949	0.8062	0.8004	0.8062	0.6365	0.6419	0.0902	0.0742	0.8708	0.9193	0.8062	0.6419	0.8062	0.6389	0.8032
0.0028	24.0	5376	1.8086	0.8085	0.8026	0.8085	0.6387	0.6429	0.0893	0.0732	0.8715	0.9200	0.8085	0.6429	0.8085	0.6405	0.8054
0.0001	25.0	5600	1.8326	0.8085	0.7988	0.8085	0.6343	0.6251	0.0934	0.0732	0.8537	0.9155	0.8085	0.6251	0.8085	0.6287	0.8028
0.0001	26.0	5824	1.8395	0.8085	0.7988	0.8085	0.6343	0.6251	0.0934	0.0732	0.8537	0.9155	0.8085	0.6251	0.8085	0.6287	0.8028
0.0003	27.0	6048	1.8816	0.8062	0.8039	0.8062	0.6439	0.6388	0.0920	0.0742	0.8621	0.9171	0.8062	0.6388	0.8062	0.6408	0.8046
0.0003	28.0	6272	1.8956	0.8062	0.8039	0.8062	0.6439	0.6388	0.0920	0.0742	0.8621	0.9171	0.8062	0.6388	0.8062	0.6408	0.8046
0.0003	29.0	6496	1.8986	0.8062	0.8039	0.8062	0.6439	0.6388	0.0920	0.0742	0.8621	0.9171	0.8062	0.6388	0.8062	0.6408	0.8046
0.0	30.0	6720	1.8999	0.8085	0.8045	0.8085	0.6455	0.6413	0.0910	0.0732	0.8622	0.9177	0.8085	0.6413	0.8085	0.6429	0.8061

Framework versions

• Transformers 4.40.2
• Pytorch 2.2.1+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1