dino-vitb16

This model is a fine-tuned version of microsoft/resnet-18 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0641
• Accuracy: 0.9856
• Precision: 0.9932
• Recall: 0.9756
• F1: 0.9843
• Tp: 1598
• Tn: 1899
• Fp: 11
• Fn: 40

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: 0.0003
• train_batch_size: 128
• eval_batch_size: 128
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 55
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1	Tp	Tn	Fp	Fn
0.4827	0.0991	11	0.6050	0.6254	0.5541	0.9664	0.7043	1583	636	1274	55
0.2848	0.1982	22	0.6828	0.6945	0.6111	0.9304	0.7377	1524	940	970	114
0.2555	0.2973	33	0.5213	0.7452	0.6491	0.9756	0.7795	1598	1046	864	40
0.2000	0.3964	44	0.1739	0.9436	0.9422	0.9353	0.9387	1532	1816	94	106
0.2370	0.4955	55	0.2988	0.8926	0.8338	0.9585	0.8918	1570	1597	313	68
0.2251	0.5946	66	0.2926	0.8929	0.8249	0.9750	0.8937	1597	1571	339	41
0.1936	0.6937	77	0.1269	0.9704	0.9936	0.9420	0.9671	1543	1900	10	95
0.1959	0.7928	88	0.1577	0.9549	0.9362	0.9683	0.9520	1586	1802	108	52
0.2104	0.8919	99	0.1320	0.9651	0.9621	0.9621	0.9621	1576	1848	62	62
0.1969	0.9910	110	0.1945	0.9453	0.9252	0.9591	0.9418	1571	1783	127	67
0.2055	1.0901	121	0.1051	0.9783	0.9962	0.9567	0.9760	1567	1904	6	71
0.1698	1.1892	132	0.1323	0.9670	0.9617	0.9670	0.9644	1584	1847	63	54
0.2012	1.2883	143	0.1589	0.9591	0.9467	0.9658	0.9562	1582	1821	89	56
0.1713	1.3874	154	0.1327	0.9583	0.9499	0.9603	0.9551	1573	1827	83	65
0.1866	1.4865	165	0.1833	0.9419	0.9877	0.8852	0.9337	1450	1892	18	188
0.1641	1.5856	176	0.1471	0.9597	0.9379	0.9774	0.9572	1601	1804	106	37
0.1749	1.6847	187	0.1130	0.9698	0.9637	0.9713	0.9675	1591	1850	60	47
0.1916	1.7838	198	0.1272	0.9746	0.9754	0.9695	0.9724	1588	1870	40	50
0.1888	1.8829	209	0.0958	0.9777	0.9779	0.9737	0.9758	1595	1874	36	43
0.1605	1.9820	220	0.0682	0.9825	0.9975	0.9646	0.9808	1580	1906	4	58
0.1522	2.0811	231	0.1065	0.9710	0.9735	0.9634	0.9684	1578	1867	43	60
0.1574	2.1802	242	0.0794	0.9789	0.9821	0.9719	0.9770	1592	1881	29	46
0.1552	2.2793	253	0.0707	0.9820	0.9931	0.9676	0.9802	1585	1899	11	53
0.1476	2.3784	264	0.0820	0.9837	0.9901	0.9744	0.9822	1596	1894	16	42
0.1670	2.4775	275	0.0803	0.9806	0.9834	0.9744	0.9788	1596	1883	27	42
0.1729	2.5766	286	0.0762	0.9825	0.9888	0.9731	0.9809	1594	1892	18	44
0.1524	2.6757	297	0.0713	0.9837	0.9907	0.9737	0.9821	1595	1895	15	43
0.1697	2.7748	308	0.0913	0.9820	0.9882	0.9725	0.9803	1593	1891	19	45
0.1429	2.8739	319	0.0824	0.9839	0.9931	0.9719	0.9824	1592	1899	11	46
0.1686	2.9730	330	0.0747	0.9831	0.9962	0.9670	0.9814	1584	1904	6	54
0.1357	3.0721	341	0.0739	0.9825	0.9888	0.9731	0.9809	1594	1892	18	44
0.1606	3.1712	352	0.0705	0.9834	0.9882	0.9756	0.9819	1598	1891	19	40
0.1697	3.2703	363	0.0640	0.9839	0.9919	0.9731	0.9824	1594	1897	13	44
0.1471	3.3694	374	0.0602	0.9873	0.9963	0.9762	0.9861	1599	1904	6	39
0.1425	3.4685	385	0.0775	0.9822	0.9828	0.9786	0.9807	1603	1882	28	35
0.1326	3.5676	396	0.0649	0.9865	0.9920	0.9786	0.9852	1603	1897	13	35
0.1342	3.6667	407	0.0559	0.9870	0.9956	0.9762	0.9858	1599	1903	7	39
0.1524	3.7658	418	0.0611	0.9865	0.9969	0.9737	0.9852	1595	1905	5	43
0.1387	3.8649	429	0.0740	0.9853	0.9913	0.9768	0.9840	1600	1896	14	38
0.1449	3.9640	440	0.0733	0.9848	0.9901	0.9768	0.9834	1600	1894	16	38
0.1605	4.0631	451	0.0756	0.9825	0.9858	0.9762	0.9810	1599	1887	23	39
0.1522	4.1622	462	0.0846	0.9800	0.9775	0.9792	0.9783	1604	1873	37	34
0.1622	4.2613	473	0.0798	0.9851	0.9877	0.9799	0.9838	1605	1890	20	33
0.1103	4.3604	484	0.0717	0.9870	0.9938	0.9780	0.9858	1602	1900	10	36
0.1381	4.4595	495	0.0676	0.9856	0.9907	0.9780	0.9843	1602	1895	15	36
0.1402	4.5586	506	0.0683	0.9845	0.9907	0.9756	0.9831	1598	1895	15	40
0.1635	4.6577	517	0.0672	0.9853	0.9913	0.9768	0.9840	1600	1896	14	38
0.1253	4.7568	528	0.0635	0.9865	0.9938	0.9768	0.9852	1600	1900	10	38
0.1409	4.8559	539	0.0630	0.9865	0.9956	0.9750	0.9852	1597	1903	7	41
0.1359	4.9550	550	0.0641	0.9856	0.9932	0.9756	0.9843	1598	1899	11	40

Framework versions

• Transformers 5.0.0
• Pytorch 2.10.0+cu128
• Datasets 4.0.0
• Tokenizers 0.22.2