dino-base-2023_10_31-demo-v5

This model is a fine-tuned version of facebook/dinov2-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0993
• F1 Micro: 0.8523
• F1 Macro: 0.7900
• Roc Auc: 0.9054
• Accuracy: 0.5712

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: 0.01
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 100

Training results

Training Loss	Epoch	Step	Validation Loss	F1 Micro	F1 Macro	Roc Auc	Accuracy
0.1686	1.0	592	0.1334	0.8014	0.7001	0.8691	0.5017
0.1255	2.0	1184	0.1214	0.8129	0.7228	0.8768	0.5180
0.1167	3.0	1776	0.1210	0.8167	0.7202	0.8714	0.5259
0.1087	4.0	2368	0.1209	0.8215	0.7468	0.8967	0.5136
0.1078	5.0	2960	0.1179	0.8135	0.7248	0.8640	0.5323
0.0992	6.0	3552	0.1169	0.8280	0.7482	0.9010	0.5224
0.0961	7.0	4144	0.1144	0.8311	0.7605	0.9004	0.5209
0.0939	8.0	4736	0.1107	0.8393	0.7503	0.9092	0.5298
0.0942	9.0	5328	0.1157	0.8249	0.7416	0.8762	0.5515
0.0922	10.0	5920	0.1072	0.8364	0.7776	0.8973	0.5481
0.0895	11.0	6512	0.1102	0.8310	0.7631	0.8890	0.5328
0.0866	12.0	7104	0.1054	0.8473	0.7701	0.9099	0.5451
0.0872	13.0	7696	0.1055	0.8454	0.7851	0.9097	0.5436
0.085	14.0	8288	0.1069	0.8422	0.7684	0.9008	0.5559
0.0854	15.0	8880	0.1106	0.8316	0.7666	0.8926	0.5456
0.0841	16.0	9472	0.1068	0.8405	0.7681	0.8922	0.5702
0.0807	17.0	10064	0.1041	0.8460	0.7814	0.9051	0.5594
0.0819	18.0	10656	0.1053	0.8431	0.7822	0.9046	0.5466
0.0801	19.0	11248	0.1081	0.8395	0.7683	0.9063	0.5436
0.0795	20.0	11840	0.1077	0.8451	0.7721	0.8980	0.5520
0.0798	21.0	12432	0.1069	0.8390	0.7721	0.8839	0.5742
0.0784	22.0	13024	0.1050	0.8442	0.7847	0.9059	0.5461
0.0775	23.0	13616	0.1065	0.8443	0.7904	0.9072	0.5476
0.0727	24.0	14208	0.1010	0.8493	0.7910	0.9051	0.5678
0.0707	25.0	14800	0.1002	0.8496	0.7877	0.9058	0.5643
0.0697	26.0	15392	0.1006	0.8489	0.7886	0.9024	0.5692
0.0699	27.0	15984	0.1005	0.8531	0.7897	0.9054	0.5702
0.0692	28.0	16576	0.1001	0.8499	0.7894	0.9059	0.5663
0.0719	29.0	17168	0.0998	0.8524	0.7854	0.9058	0.5737
0.0686	30.0	17760	0.1006	0.8503	0.7897	0.9033	0.5663
0.0692	31.0	18352	0.1000	0.8519	0.7928	0.9055	0.5717
0.0707	32.0	18944	0.1000	0.8517	0.7862	0.9056	0.5737
0.0695	33.0	19536	0.1002	0.8517	0.7850	0.9012	0.5781
0.069	34.0	20128	0.1008	0.8477	0.7849	0.9003	0.5658
0.0686	35.0	20720	0.1004	0.8523	0.7866	0.9009	0.5732
0.0688	36.0	21312	0.0994	0.8517	0.7902	0.9058	0.5673
0.0688	37.0	21904	0.0994	0.8523	0.7900	0.9048	0.5732
0.0677	38.0	22496	0.0994	0.8520	0.7905	0.9051	0.5697
0.0678	39.0	23088	0.0995	0.8516	0.7911	0.9035	0.5747
0.068	40.0	23680	0.0994	0.8520	0.7888	0.9039	0.5712
0.0679	41.0	24272	0.0994	0.8535	0.7908	0.9056	0.5757
0.0682	42.0	24864	0.0993	0.8517	0.7883	0.9054	0.5707
0.0677	43.0	25456	0.0994	0.8516	0.7908	0.9052	0.5707
0.0678	44.0	26048	0.0995	0.8518	0.7916	0.9066	0.5673
0.0677	45.0	26640	0.0993	0.8519	0.7886	0.9054	0.5702
0.0684	46.0	27232	0.0995	0.8519	0.7909	0.9060	0.5697
0.0675	47.0	27824	0.0994	0.8524	0.7908	0.9048	0.5757
0.067	48.0	28416	0.0995	0.8521	0.7893	0.9044	0.5717
0.0675	49.0	29008	0.0994	0.8524	0.7902	0.9056	0.5707
0.0674	50.0	29600	0.0994	0.8517	0.7893	0.9051	0.5692
0.0679	51.0	30192	0.0993	0.8519	0.7898	0.9052	0.5697
0.0667	52.0	30784	0.0993	0.8523	0.7900	0.9054	0.5712

Framework versions

• Transformers 4.34.1
• Pytorch 2.1.0+cu118
• Datasets 2.14.5
• Tokenizers 0.14.1