vit-tiny_tobacco3482_og_simkd_

This model is a fine-tuned version of WinKawaks/vit-tiny-patch16-224 on the None dataset. It achieves the following results on the evaluation set:

Loss: 212.9255
Accuracy: 0.825
Brier Loss: 0.2955
Nll: 1.3247
F1 Micro: 0.825
F1 Macro: 0.8027
Ece: 0.1660
Aurc: 0.0554

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: 0.0001
train_batch_size: 32
eval_batch_size: 32
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
lr_scheduler_warmup_ratio: 0.1
num_epochs: 100

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Brier Loss	Nll	F1 Micro	F1 Macro	Ece	Aurc
No log	1.0	25	219.5859	0.225	0.8919	5.3905	0.225	0.1403	0.2467	0.7026
No log	2.0	50	218.7605	0.385	0.8149	2.3886	0.3850	0.3040	0.3416	0.4009
No log	3.0	75	217.4839	0.49	0.6311	1.8415	0.49	0.3775	0.2502	0.2789
No log	4.0	100	216.7303	0.605	0.5512	1.9525	0.605	0.4738	0.2423	0.1956
No log	5.0	125	215.7727	0.625	0.4792	1.8491	0.625	0.5148	0.2037	0.1503
No log	6.0	150	215.9141	0.595	0.5723	2.5766	0.595	0.4836	0.2297	0.2039
No log	7.0	175	215.2466	0.715	0.4245	1.8434	0.715	0.5962	0.2236	0.1201
No log	8.0	200	215.0580	0.72	0.3878	1.8760	0.72	0.6108	0.2068	0.0879
No log	9.0	225	214.9401	0.745	0.3919	1.7852	0.745	0.7183	0.1855	0.0936
No log	10.0	250	214.6627	0.66	0.4840	2.1588	0.66	0.6144	0.2371	0.1261
No log	11.0	275	214.9265	0.69	0.5027	1.7910	0.69	0.6599	0.2514	0.1279
No log	12.0	300	214.4700	0.78	0.3372	1.7013	0.78	0.7459	0.1740	0.0741
No log	13.0	325	214.5191	0.805	0.3164	1.5824	0.805	0.7689	0.1662	0.0689
No log	14.0	350	214.3915	0.8	0.3278	1.5290	0.8000	0.7742	0.1716	0.0675
No log	15.0	375	214.2643	0.8	0.3315	1.5964	0.8000	0.7697	0.1491	0.0876
No log	16.0	400	214.2815	0.78	0.3751	1.8388	0.78	0.7670	0.1944	0.0836
No log	17.0	425	214.0954	0.78	0.3505	1.4905	0.78	0.7521	0.1832	0.0662
No log	18.0	450	214.1399	0.785	0.3392	1.2883	0.785	0.7662	0.1760	0.0749
No log	19.0	475	214.1986	0.81	0.3229	1.6305	0.81	0.7984	0.1754	0.0777
220.3818	20.0	500	214.1931	0.815	0.2994	1.5637	0.815	0.7940	0.1520	0.0606
220.3818	21.0	525	213.9438	0.815	0.3066	1.4756	0.815	0.7983	0.1616	0.0678
220.3818	22.0	550	213.9014	0.8	0.3485	1.9629	0.8000	0.7885	0.1838	0.0727
220.3818	23.0	575	214.0186	0.83	0.2863	1.6314	0.83	0.8107	0.1516	0.0574
220.3818	24.0	600	213.8764	0.805	0.3323	1.4873	0.805	0.7903	0.1716	0.0726
220.3818	25.0	625	214.0043	0.81	0.3094	1.5562	0.81	0.7764	0.1625	0.0598
220.3818	26.0	650	213.6884	0.825	0.3165	1.6150	0.825	0.8038	0.1667	0.0755
220.3818	27.0	675	213.7763	0.81	0.3164	1.5526	0.81	0.7904	0.1767	0.0747
220.3818	28.0	700	213.9658	0.825	0.2996	1.7947	0.825	0.8149	0.1686	0.0651
220.3818	29.0	725	213.7030	0.815	0.3155	1.3772	0.815	0.8024	0.1616	0.0613
220.3818	30.0	750	213.7211	0.805	0.3421	1.5621	0.805	0.7986	0.1794	0.0624
220.3818	31.0	775	213.6852	0.815	0.3094	1.5177	0.815	0.7872	0.1500	0.0735
220.3818	32.0	800	213.6889	0.785	0.3345	1.4134	0.785	0.7652	0.1669	0.0563
220.3818	33.0	825	213.6302	0.805	0.3298	1.8630	0.805	0.7865	0.1689	0.0693
220.3818	34.0	850	213.6116	0.83	0.2890	1.5365	0.83	0.8033	0.1555	0.0661
220.3818	35.0	875	213.6136	0.805	0.3026	1.1797	0.805	0.7744	0.1551	0.0477
220.3818	36.0	900	213.5340	0.815	0.3008	1.6963	0.815	0.7938	0.1531	0.0687
220.3818	37.0	925	213.5380	0.84	0.2968	1.3848	0.8400	0.8396	0.1631	0.0577
220.3818	38.0	950	213.5322	0.83	0.3016	1.5511	0.83	0.8133	0.1647	0.0599
220.3818	39.0	975	213.4971	0.82	0.3155	1.4175	0.82	0.7999	0.1571	0.0535
217.7955	40.0	1000	213.4139	0.825	0.3103	1.6359	0.825	0.8043	0.1749	0.0683
217.7955	41.0	1025	213.4513	0.83	0.3002	1.5369	0.83	0.8139	0.1580	0.0606
217.7955	42.0	1050	213.4196	0.8	0.3251	1.3570	0.8000	0.7779	0.1745	0.0608
217.7955	43.0	1075	213.3506	0.815	0.3142	1.3579	0.815	0.7988	0.1724	0.0563
217.7955	44.0	1100	213.3151	0.805	0.3217	1.3796	0.805	0.7820	0.1748	0.0580
217.7955	45.0	1125	213.3202	0.825	0.3114	1.4198	0.825	0.8090	0.1568	0.0614
217.7955	46.0	1150	213.3313	0.805	0.3203	1.3750	0.805	0.7860	0.1667	0.0563
217.7955	47.0	1175	213.3293	0.835	0.2910	1.3909	0.835	0.8190	0.1515	0.0576
217.7955	48.0	1200	213.2646	0.825	0.2916	1.3674	0.825	0.8022	0.1526	0.0577
217.7955	49.0	1225	213.2620	0.83	0.3137	1.4579	0.83	0.8101	0.1634	0.0565
217.7955	50.0	1250	213.2164	0.815	0.3087	1.2599	0.815	0.7917	0.1618	0.0573
217.7955	51.0	1275	213.2495	0.795	0.3155	1.2060	0.795	0.7648	0.1712	0.0573
217.7955	52.0	1300	213.2231	0.82	0.3232	1.4463	0.82	0.8029	0.1658	0.0544
217.7955	53.0	1325	213.2242	0.83	0.2891	1.3586	0.83	0.8056	0.1427	0.0520
217.7955	54.0	1350	213.2049	0.83	0.2959	1.2968	0.83	0.8063	0.1573	0.0523
217.7955	55.0	1375	213.1500	0.84	0.2844	1.4084	0.8400	0.8137	0.1490	0.0570
217.7955	56.0	1400	213.1851	0.815	0.3097	1.4005	0.815	0.7892	0.1605	0.0605
217.7955	57.0	1425	213.1577	0.805	0.3130	1.2482	0.805	0.7805	0.1635	0.0550
217.7955	58.0	1450	213.1812	0.835	0.2943	1.3191	0.835	0.8047	0.1584	0.0550
217.7955	59.0	1475	213.0962	0.82	0.3102	1.3163	0.82	0.7906	0.1672	0.0513
217.1347	60.0	1500	213.1257	0.835	0.2876	1.3165	0.835	0.8122	0.1433	0.0562
217.1347	61.0	1525	213.1092	0.83	0.2896	1.3196	0.83	0.8061	0.1533	0.0549
217.1347	62.0	1550	213.0827	0.815	0.3072	1.3053	0.815	0.7925	0.1546	0.0523
217.1347	63.0	1575	213.0972	0.825	0.2965	1.2148	0.825	0.8044	0.1472	0.0507
217.1347	64.0	1600	213.0849	0.83	0.2923	1.3196	0.83	0.8106	0.1637	0.0551
217.1347	65.0	1625	213.0354	0.815	0.3158	1.2749	0.815	0.7867	0.1675	0.0515
217.1347	66.0	1650	213.0493	0.825	0.2993	1.3046	0.825	0.7949	0.1636	0.0558
217.1347	67.0	1675	212.9869	0.815	0.3081	1.3846	0.815	0.7868	0.1579	0.0583
217.1347	68.0	1700	213.0629	0.84	0.2915	1.3902	0.8400	0.8046	0.1590	0.0540
217.1347	69.0	1725	213.0391	0.825	0.3068	1.3801	0.825	0.8015	0.1553	0.0531
217.1347	70.0	1750	212.9991	0.835	0.2864	1.3331	0.835	0.8097	0.1531	0.0562
217.1347	71.0	1775	213.0157	0.83	0.2897	1.2788	0.83	0.8002	0.1584	0.0551
217.1347	72.0	1800	212.9134	0.82	0.3051	1.3131	0.82	0.7960	0.1690	0.0522
217.1347	73.0	1825	213.0014	0.825	0.2926	1.3111	0.825	0.8040	0.1390	0.0574
217.1347	74.0	1850	212.9525	0.82	0.2985	1.3181	0.82	0.7962	0.1579	0.0543
217.1347	75.0	1875	212.9581	0.815	0.3024	1.2835	0.815	0.7810	0.1648	0.0504
217.1347	76.0	1900	213.0073	0.835	0.2970	1.3745	0.835	0.8095	0.1597	0.0579
217.1347	77.0	1925	213.0066	0.805	0.3046	1.3071	0.805	0.7783	0.1502	0.0547
217.1347	78.0	1950	212.9872	0.82	0.3018	1.4088	0.82	0.7928	0.1527	0.0527
217.1347	79.0	1975	212.9629	0.82	0.3024	1.3665	0.82	0.8012	0.1626	0.0551
216.794	80.0	2000	212.9545	0.825	0.3080	1.3609	0.825	0.8062	0.1652	0.0541
216.794	81.0	2025	212.9253	0.825	0.3077	1.3779	0.825	0.8044	0.1662	0.0547
216.794	82.0	2050	212.9501	0.82	0.3024	1.3636	0.82	0.7928	0.1677	0.0553
216.794	83.0	2075	212.9160	0.81	0.3055	1.3686	0.81	0.7786	0.1624	0.0578
216.794	84.0	2100	212.9532	0.84	0.2914	1.4589	0.8400	0.8129	0.1482	0.0510
216.794	85.0	2125	212.9397	0.825	0.3067	1.3768	0.825	0.7981	0.1653	0.0537
216.794	86.0	2150	212.8927	0.83	0.2980	1.3825	0.83	0.8118	0.1662	0.0560
216.794	87.0	2175	212.8856	0.825	0.3004	1.4017	0.825	0.8063	0.1595	0.0555
216.794	88.0	2200	212.9423	0.82	0.3033	1.3619	0.82	0.8012	0.1539	0.0517
216.794	89.0	2225	212.8776	0.84	0.2922	1.3845	0.8400	0.8176	0.1555	0.0537
216.794	90.0	2250	212.9439	0.82	0.3011	1.3923	0.82	0.8012	0.1526	0.0535
216.794	91.0	2275	212.8640	0.815	0.3006	1.3680	0.815	0.7920	0.1443	0.0548
216.794	92.0	2300	212.8850	0.825	0.2940	1.3317	0.825	0.8087	0.1466	0.0548
216.794	93.0	2325	212.8843	0.825	0.3024	1.3848	0.825	0.8027	0.1688	0.0529
216.794	94.0	2350	212.9464	0.825	0.3013	1.3634	0.825	0.8027	0.1607	0.0515
216.794	95.0	2375	212.9154	0.825	0.3001	1.3262	0.825	0.8087	0.1604	0.0539
216.794	96.0	2400	212.9131	0.825	0.2984	1.3121	0.825	0.8027	0.1568	0.0538
216.794	97.0	2425	212.8888	0.82	0.3003	1.3975	0.82	0.8012	0.1614	0.0527
216.794	98.0	2450	212.8581	0.825	0.2989	1.2911	0.825	0.8027	0.1673	0.0542
216.794	99.0	2475	212.9360	0.83	0.2983	1.3929	0.83	0.8118	0.1616	0.0534
216.6176	100.0	2500	212.9255	0.825	0.2955	1.3247	0.825	0.8027	0.1660	0.0554

Framework versions

Transformers 4.26.1
Pytorch 1.13.1.post200
Datasets 2.9.0
Tokenizers 0.13.2

jordyvl/vit-tiny_tobacco3482_og_simkd_

vit-tiny_tobacco3482_og_simkd_

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions