xshubhamx/InLegalBERT-lora-no-grad

legal-bert-lora-no-grad

This model is a fine-tuned version of law-ai/InLegalBERT on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.5075
• Accuracy: 0.8280
• Precision: 0.8290
• Recall: 0.8280
• Precision Macro: 0.7852
• Recall Macro: 0.7756
• Macro Fpr: 0.0151
• Weighted Fpr: 0.0145
• Weighted Specificity: 0.9775
• Macro Specificity: 0.9871
• Weighted Sensitivity: 0.8288
• Macro Sensitivity: 0.7756
• F1 Micro: 0.8288
• F1 Macro: 0.7761
• F1 Weighted: 0.8279

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
1.6412	1.0	643	0.7925	0.7514	0.7190	0.7514	0.4123	0.4707	0.0237	0.0231	0.9699	0.9814	0.7514	0.4707	0.7514	0.4277	0.7283
0.7481	2.0	1286	0.6772	0.7901	0.7726	0.7901	0.5958	0.6252	0.0192	0.0186	0.9741	0.9843	0.7901	0.6252	0.7901	0.5998	0.7769
0.6465	3.0	1929	0.6500	0.8048	0.7931	0.8048	0.6216	0.6414	0.0176	0.0170	0.9764	0.9854	0.8048	0.6414	0.8048	0.6110	0.7904
0.4707	4.0	2572	0.6704	0.8095	0.8008	0.8095	0.6322	0.6689	0.0173	0.0165	0.9745	0.9856	0.8095	0.6689	0.8095	0.6425	0.8018
0.4021	5.0	3215	0.7320	0.8280	0.8269	0.8280	0.7782	0.7573	0.0154	0.0146	0.9765	0.9870	0.8280	0.7573	0.8280	0.7571	0.8219
0.3627	6.0	3858	0.6892	0.8242	0.8227	0.8242	0.7431	0.7365	0.0156	0.0150	0.9768	0.9867	0.8242	0.7365	0.8242	0.7374	0.8223
0.2866	7.0	4501	0.8756	0.8180	0.8171	0.8180	0.7748	0.7410	0.0166	0.0156	0.9718	0.9860	0.8180	0.7410	0.8180	0.7444	0.8122
0.2639	8.0	5144	0.8580	0.8265	0.8259	0.8265	0.7989	0.7428	0.0155	0.0148	0.9756	0.9868	0.8265	0.7428	0.8265	0.7480	0.8217
0.2295	9.0	5787	0.9366	0.8257	0.8231	0.8257	0.7725	0.7465	0.0155	0.0149	0.9762	0.9868	0.8257	0.7465	0.8257	0.7524	0.8223
0.195	10.0	6430	0.9685	0.8273	0.8236	0.8273	0.7595	0.7515	0.0153	0.0147	0.9767	0.9869	0.8273	0.7515	0.8273	0.7528	0.8241
0.1617	11.0	7073	1.0406	0.8311	0.8263	0.8311	0.7615	0.7552	0.0149	0.0143	0.9776	0.9872	0.8311	0.7552	0.8311	0.7543	0.8265
0.1421	12.0	7716	1.0713	0.8319	0.8276	0.8319	0.7626	0.7533	0.0148	0.0142	0.9773	0.9873	0.8319	0.7533	0.8319	0.7546	0.8287
0.1184	13.0	8359	1.1125	0.8257	0.8209	0.8257	0.7569	0.7504	0.0155	0.0149	0.9765	0.9868	0.8257	0.7504	0.8257	0.7510	0.8219
0.1017	14.0	9002	1.1926	0.8211	0.8215	0.8211	0.7675	0.7815	0.0159	0.0153	0.9776	0.9866	0.8211	0.7815	0.8211	0.7727	0.8196
0.0752	15.0	9645	1.2508	0.8164	0.8121	0.8164	0.7479	0.7377	0.0164	0.0158	0.9753	0.9861	0.8164	0.7377	0.8164	0.7402	0.8133
0.0787	16.0	10288	1.3247	0.8218	0.8199	0.8218	0.8034	0.7585	0.0160	0.0152	0.9752	0.9865	0.8218	0.7585	0.8218	0.7698	0.8188
0.0668	17.0	10931	1.3497	0.8211	0.8201	0.8211	0.7500	0.7487	0.0158	0.0153	0.9778	0.9866	0.8211	0.7487	0.8211	0.7468	0.8198
0.0471	18.0	11574	1.4278	0.8164	0.8174	0.8164	0.7672	0.7670	0.0165	0.0158	0.9759	0.9862	0.8164	0.7670	0.8164	0.7644	0.8159
0.0492	19.0	12217	1.4784	0.8180	0.8178	0.8180	0.7631	0.7431	0.0162	0.0156	0.9763	0.9863	0.8180	0.7431	0.8180	0.7453	0.8156
0.0368	20.0	12860	1.4747	0.8195	0.8183	0.8195	0.7729	0.7568	0.0161	0.0155	0.9760	0.9864	0.8195	0.7568	0.8195	0.7622	0.8180
0.0329	21.0	13503	1.5075	0.8280	0.8290	0.8280	0.7825	0.7845	0.0152	0.0146	0.9782	0.9871	0.8280	0.7845	0.8280	0.7798	0.8268
0.0266	22.0	14146	1.4783	0.8273	0.8262	0.8273	0.7780	0.7612	0.0153	0.0147	0.9779	0.9870	0.8273	0.7612	0.8273	0.7651	0.8247
0.0302	23.0	14789	1.5281	0.8234	0.8246	0.8234	0.7745	0.7699	0.0158	0.0151	0.9760	0.9866	0.8234	0.7699	0.8234	0.7679	0.8224
0.0207	24.0	15432	1.5475	0.8265	0.8262	0.8265	0.7809	0.7727	0.0155	0.0148	0.9768	0.9869	0.8265	0.7727	0.8265	0.7721	0.8248
0.0168	25.0	16075	1.5237	0.8242	0.8237	0.8242	0.7726	0.7619	0.0155	0.0150	0.9775	0.9868	0.8242	0.7619	0.8242	0.7629	0.8231
0.0167	26.0	16718	1.5815	0.8234	0.8255	0.8234	0.7766	0.7728	0.0156	0.0151	0.9775	0.9867	0.8234	0.7728	0.8234	0.7707	0.8232
0.0127	27.0	17361	1.6010	0.8218	0.8228	0.8218	0.7790	0.7716	0.0158	0.0152	0.9769	0.9866	0.8218	0.7716	0.8218	0.7709	0.8211
0.0094	28.0	18004	1.5774	0.8265	0.8269	0.8265	0.7788	0.7739	0.0153	0.0148	0.9778	0.9870	0.8265	0.7739	0.8265	0.7728	0.8258
0.0063	29.0	18647	1.5894	0.8304	0.8306	0.8304	0.7825	0.7764	0.0150	0.0144	0.9779	0.9872	0.8304	0.7764	0.8304	0.7759	0.8296
0.0126	30.0	19290	1.5927	0.8288	0.8291	0.8288	0.7852	0.7756	0.0151	0.0145	0.9775	0.9871	0.8288	0.7756	0.8288	0.7761	0.8279

Framework versions

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