legal-bert-lora-no-grad

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: 1.4787
• Accuracy: 0.8273
• Precision: 0.8270
• Recall: 0.8273
• Precision Macro: 0.7941
• Recall Macro: 0.7774
• Macro Fpr: 0.0157
• Weighted Fpr: 0.0151
• Weighted Specificity: 0.9767
• Macro Specificity: 0.9867
• Weighted Sensitivity: 0.8234
• Macro Sensitivity: 0.7774
• F1 Micro: 0.8234
• F1 Macro: 0.7824
• F1 Weighted: 0.8225

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

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
1.5473	1.0	643	0.8485	0.7173	0.6897	0.7173	0.3892	0.4422	0.0278	0.0274	0.9658	0.9789	0.7173	0.4422	0.7173	0.3920	0.6869
0.7816	2.0	1286	0.7113	0.7545	0.7492	0.7545	0.5282	0.5054	0.0231	0.0227	0.9711	0.9817	0.7545	0.5054	0.7545	0.4751	0.7304
0.6956	3.0	1929	0.6460	0.7986	0.7793	0.7986	0.5436	0.5701	0.0184	0.0177	0.9741	0.9848	0.7986	0.5701	0.7986	0.5439	0.7833
0.4942	4.0	2572	0.6430	0.8110	0.8014	0.8110	0.6315	0.6670	0.0169	0.0164	0.9763	0.9858	0.8110	0.6670	0.8110	0.6413	0.8037
0.4088	5.0	3215	0.7148	0.8319	0.8301	0.8319	0.7951	0.7444	0.0150	0.0142	0.9758	0.9872	0.8319	0.7444	0.8319	0.7459	0.8254
0.3722	6.0	3858	0.7203	0.8319	0.8282	0.8319	0.7594	0.7604	0.0149	0.0142	0.9779	0.9873	0.8319	0.7604	0.8319	0.7582	0.8285
0.3088	7.0	4501	0.7796	0.8218	0.8283	0.8218	0.7927	0.7436	0.0158	0.0152	0.9765	0.9866	0.8218	0.7436	0.8218	0.7486	0.8203
0.245	8.0	5144	0.8732	0.8187	0.8171	0.8187	0.7696	0.7403	0.0163	0.0156	0.9752	0.9863	0.8187	0.7403	0.8187	0.7409	0.8135
0.2331	9.0	5787	0.8710	0.8265	0.8280	0.8265	0.7595	0.7521	0.0152	0.0148	0.9776	0.9869	0.8265	0.7521	0.8265	0.7530	0.8261
0.1878	10.0	6430	0.9866	0.8257	0.8261	0.8257	0.7672	0.7604	0.0153	0.0149	0.9788	0.9870	0.8257	0.7604	0.8257	0.7588	0.8240
0.1627	11.0	7073	1.0530	0.8257	0.8269	0.8257	0.7706	0.7711	0.0154	0.0149	0.9787	0.9870	0.8257	0.7711	0.8257	0.7675	0.8234
0.1301	12.0	7716	1.1042	0.8265	0.8246	0.8265	0.7633	0.7587	0.0155	0.0148	0.9763	0.9869	0.8265	0.7587	0.8265	0.7576	0.8228
0.1291	13.0	8359	1.1461	0.8234	0.8215	0.8234	0.7582	0.7613	0.0157	0.0151	0.9768	0.9867	0.8234	0.7613	0.8234	0.7581	0.8211
0.11	14.0	9002	1.1837	0.8226	0.8182	0.8226	0.7716	0.7576	0.0159	0.0152	0.9756	0.9865	0.8226	0.7576	0.8226	0.7627	0.8195
0.0863	15.0	9645	1.2020	0.8218	0.8185	0.8218	0.7616	0.7458	0.0160	0.0152	0.9754	0.9865	0.8218	0.7458	0.8218	0.7498	0.8183
0.0735	16.0	10288	1.2491	0.8187	0.8160	0.8187	0.7620	0.7464	0.0162	0.0156	0.9755	0.9863	0.8187	0.7464	0.8187	0.7517	0.8168
0.0802	17.0	10931	1.3288	0.8164	0.8165	0.8164	0.7471	0.7531	0.0164	0.0158	0.9765	0.9862	0.8164	0.7531	0.8164	0.7483	0.8152
0.0525	18.0	11574	1.3620	0.8133	0.8127	0.8133	0.7557	0.7478	0.0168	0.0161	0.9754	0.9859	0.8133	0.7478	0.8133	0.7486	0.8106
0.0474	19.0	12217	1.3783	0.8187	0.8220	0.8187	0.8154	0.7972	0.0162	0.0156	0.9755	0.9863	0.8187	0.7972	0.8187	0.8024	0.8187
0.0315	20.0	12860	1.4004	0.8226	0.8266	0.8226	0.8011	0.7910	0.0157	0.0152	0.9778	0.9867	0.8226	0.7910	0.8226	0.7907	0.8231
0.0325	21.0	13503	1.4683	0.8187	0.8198	0.8187	0.8030	0.7876	0.0161	0.0156	0.9765	0.9863	0.8187	0.7876	0.8187	0.7921	0.8182
0.0192	22.0	14146	1.4677	0.8249	0.8224	0.8249	0.7598	0.7482	0.0155	0.0149	0.9768	0.9868	0.8249	0.7482	0.8249	0.7505	0.8225
0.0235	23.0	14789	1.4610	0.8211	0.8221	0.8211	0.8090	0.7941	0.0160	0.0153	0.9765	0.9865	0.8211	0.7941	0.8211	0.7982	0.8200
0.0142	24.0	15432	1.4787	0.8273	0.8270	0.8273	0.8179	0.7977	0.0153	0.0147	0.9774	0.9870	0.8273	0.7977	0.8273	0.8054	0.8260
0.0172	25.0	16075	1.5374	0.8211	0.8231	0.8211	0.7830	0.7711	0.0159	0.0153	0.9767	0.9865	0.8211	0.7711	0.8211	0.7737	0.8212
0.0097	26.0	16718	1.5153	0.8242	0.8262	0.8242	0.8105	0.7894	0.0156	0.0150	0.9772	0.9868	0.8242	0.7894	0.8242	0.7954	0.8240
0.0109	27.0	17361	1.5565	0.8218	0.8237	0.8218	0.7876	0.7848	0.0158	0.0152	0.9770	0.9866	0.8218	0.7848	0.8218	0.7832	0.8214
0.0076	28.0	18004	1.5574	0.8226	0.8247	0.8226	0.7882	0.7843	0.0157	0.0152	0.9771	0.9867	0.8226	0.7843	0.8226	0.7836	0.8229
0.006	29.0	18647	1.5721	0.8234	0.8235	0.8234	0.7868	0.7771	0.0157	0.0151	0.9769	0.9867	0.8234	0.7771	0.8234	0.7788	0.8224
0.0035	30.0	19290	1.5803	0.8234	0.8234	0.8234	0.7941	0.7774	0.0157	0.0151	0.9767	0.9867	0.8234	0.7774	0.8234	0.7824	0.8225

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu121
• Datasets 2.18.0
• Tokenizers 0.15.1