mi-1.2-model

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

• Loss: 1.7264
• Accuracy: 0.58

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: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.6501	0.04	10	1.6095	0.235
1.655	0.08	20	1.5876	0.23
1.6465	0.12	30	1.5874	0.305
1.6577	0.16	40	1.6006	0.2325
1.5666	0.2	50	1.5611	0.245
1.5667	0.24	60	1.4245	0.44
1.4837	0.28	70	1.2916	0.4175
1.2603	0.32	80	1.3869	0.3925
1.2865	0.36	90	1.4055	0.3475
1.4037	0.4	100	1.3934	0.32
1.3201	0.44	110	1.4511	0.4125
1.3977	0.48	120	1.2251	0.44
1.1444	0.52	130	1.1517	0.5175
1.1627	0.56	140	1.1211	0.5225
1.21	0.6	150	1.1336	0.53
1.2211	0.64	160	1.4186	0.4
1.2985	0.68	170	1.1251	0.4725
1.1856	0.72	180	1.1138	0.5075
1.1027	0.76	190	1.0810	0.5075
1.0998	0.8	200	1.1034	0.5225
1.2546	0.84	210	1.1205	0.4925
1.0265	0.88	220	1.1996	0.4925
1.0898	0.92	230	1.1002	0.515
1.19	0.96	240	1.0805	0.4925
1.1456	1.0	250	1.0509	0.525
0.9265	1.04	260	1.1092	0.51
0.8554	1.08	270	1.0098	0.5325
0.8695	1.12	280	1.0991	0.4975
0.8505	1.16	290	1.0827	0.5075
0.8892	1.2	300	1.1195	0.52
0.8982	1.24	310	1.0691	0.51
0.9301	1.28	320	1.0236	0.545
1.052	1.32	330	1.0296	0.535
0.8072	1.3600	340	1.0227	0.55
0.8822	1.4	350	1.0494	0.53
1.1561	1.44	360	1.2036	0.4925
0.9526	1.48	370	1.0443	0.56
0.9916	1.52	380	1.0378	0.555
1.0388	1.56	390	1.0920	0.5375
0.9326	1.6	400	1.0510	0.5375
0.8453	1.6400	410	1.1247	0.5025
1.03	1.6800	420	1.0281	0.565
0.971	1.72	430	1.0322	0.54
0.941	1.76	440	0.9858	0.565
0.8615	1.8	450	0.9793	0.555
0.8815	1.8400	460	0.9778	0.56
0.7658	1.88	470	0.9760	0.56
1.0073	1.92	480	1.0747	0.5175
0.8929	1.96	490	0.9910	0.565
0.9089	2.0	500	1.0512	0.535
0.5102	2.04	510	1.0545	0.555
0.6748	2.08	520	1.1621	0.5175
0.5222	2.12	530	1.1038	0.5575
0.7978	2.16	540	1.1728	0.53
0.6749	2.2	550	1.1029	0.5475
0.6621	2.24	560	1.0977	0.5425
0.6808	2.2800	570	1.1776	0.545
0.5728	2.32	580	1.1747	0.5325
0.75	2.36	590	1.1707	0.5275
0.6622	2.4	600	1.1082	0.555
0.6008	2.44	610	1.0922	0.57
0.6491	2.48	620	1.1375	0.545
0.5876	2.52	630	1.0614	0.5675
0.5326	2.56	640	1.0460	0.58
0.4901	2.6	650	1.0864	0.58
0.6151	2.64	660	1.1919	0.58
0.6478	2.68	670	1.1301	0.5575
0.4841	2.7200	680	1.1451	0.58
0.6365	2.76	690	1.0701	0.575
0.5284	2.8	700	1.1674	0.5325
0.6506	2.84	710	1.1016	0.55
0.6446	2.88	720	1.1340	0.57
0.5193	2.92	730	1.1692	0.525
0.6129	2.96	740	1.1717	0.5325
0.6013	3.0	750	1.1374	0.55
0.3392	3.04	760	1.2702	0.515
0.3188	3.08	770	1.2584	0.515
0.3272	3.12	780	1.3520	0.5225
0.341	3.16	790	1.2752	0.5575
0.3826	3.2	800	1.3126	0.55
0.3062	3.24	810	1.4909	0.52
0.2657	3.2800	820	1.3804	0.5575
0.4609	3.32	830	1.3712	0.5625
0.3388	3.36	840	1.4701	0.5275
0.3007	3.4	850	1.3373	0.57
0.2732	3.44	860	1.3699	0.575
0.4551	3.48	870	1.3874	0.555
0.3048	3.52	880	1.4913	0.5625
0.4104	3.56	890	1.4586	0.565
0.2633	3.6	900	1.4353	0.565
0.4435	3.64	910	1.5246	0.555
0.282	3.68	920	1.6866	0.5275
0.5918	3.7200	930	1.5193	0.5525
0.315	3.76	940	1.4276	0.565
0.1276	3.8	950	1.4411	0.5625
0.3389	3.84	960	1.5420	0.5625
0.3248	3.88	970	1.4492	0.575
0.3051	3.92	980	1.4321	0.5925
0.3363	3.96	990	1.4374	0.5825
0.4602	4.0	1000	1.4581	0.57
0.1582	4.04	1010	1.4434	0.5675
0.2344	4.08	1020	1.4551	0.5975
0.2646	4.12	1030	1.4999	0.59
0.1948	4.16	1040	1.5550	0.5625
0.3058	4.2	1050	1.5955	0.5775
0.1569	4.24	1060	1.5721	0.575
0.1777	4.28	1070	1.6241	0.56
0.1256	4.32	1080	1.5711	0.575
0.2467	4.36	1090	1.5735	0.59
0.1964	4.4	1100	1.5924	0.585
0.0578	4.44	1110	1.6353	0.585
0.1358	4.48	1120	1.6710	0.5775
0.174	4.52	1130	1.6733	0.5725
0.2022	4.5600	1140	1.6658	0.585
0.028	4.6	1150	1.6708	0.585
0.1222	4.64	1160	1.6989	0.5875
0.2295	4.68	1170	1.7131	0.5825
0.374	4.72	1180	1.7197	0.5725
0.1342	4.76	1190	1.7237	0.575
0.079	4.8	1200	1.7267	0.58
0.154	4.84	1210	1.7204	0.585
0.0403	4.88	1220	1.7183	0.58
0.1964	4.92	1230	1.7253	0.5775
0.1297	4.96	1240	1.7252	0.5775
0.0834	5.0	1250	1.7264	0.58

Framework versions

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