vit-small_tobacco3482_kd_NKD_t1.0_g1.5

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

Loss: 3.9399
Accuracy: 0.82
Brier Loss: 0.3024
Nll: 1.1952
F1 Micro: 0.82
F1 Macro: 0.7964
Ece: 0.1494
Aurc: 0.0548

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: 128
eval_batch_size: 128
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	7	4.9730	0.205	0.8886	5.4337	0.205	0.1356	0.2736	0.7527
No log	2.0	14	4.6039	0.355	0.8122	3.5880	0.3550	0.2120	0.3197	0.5132
No log	3.0	21	4.2754	0.515	0.7054	2.0283	0.515	0.4046	0.3275	0.2966
No log	4.0	28	4.0263	0.6	0.5799	1.5709	0.6	0.5157	0.3062	0.1890
No log	5.0	35	3.8749	0.725	0.4857	1.5338	0.7250	0.6949	0.3181	0.1194
No log	6.0	42	3.7023	0.765	0.3925	1.2926	0.765	0.6948	0.2394	0.0908
No log	7.0	49	3.7728	0.78	0.3668	1.3007	0.78	0.7355	0.2478	0.0754
No log	8.0	56	3.7328	0.785	0.3459	1.2487	0.785	0.7501	0.2261	0.0804
No log	9.0	63	3.7092	0.77	0.3289	1.0921	0.7700	0.7672	0.2019	0.0767
No log	10.0	70	3.6273	0.795	0.3150	1.0342	0.795	0.7690	0.1927	0.0716
No log	11.0	77	3.5677	0.83	0.2754	1.3837	0.83	0.7933	0.1697	0.0532
No log	12.0	84	3.5668	0.815	0.2816	1.1304	0.815	0.7934	0.1563	0.0617
No log	13.0	91	3.6080	0.83	0.2723	0.9515	0.83	0.8088	0.1648	0.0543
No log	14.0	98	3.6095	0.815	0.3050	1.2020	0.815	0.8207	0.1523	0.0633
No log	15.0	105	3.6685	0.805	0.3060	1.2725	0.805	0.7920	0.1618	0.0676
No log	16.0	112	3.5523	0.825	0.2832	0.9447	0.825	0.8163	0.1614	0.0569
No log	17.0	119	3.5294	0.805	0.2752	0.9918	0.805	0.7636	0.1537	0.0549
No log	18.0	126	3.5382	0.8	0.2870	1.2294	0.8000	0.7885	0.1603	0.0583
No log	19.0	133	3.5541	0.82	0.2905	1.2181	0.82	0.8204	0.1400	0.0618
No log	20.0	140	3.4717	0.835	0.2606	1.1119	0.835	0.8146	0.1382	0.0531
No log	21.0	147	3.6074	0.79	0.3099	1.2144	0.79	0.7771	0.1419	0.0599
No log	22.0	154	3.5448	0.805	0.2868	1.2075	0.805	0.7761	0.1439	0.0581
No log	23.0	161	3.6070	0.805	0.3057	1.2908	0.805	0.7831	0.1393	0.0627
No log	24.0	168	3.5289	0.81	0.2716	1.1844	0.81	0.7879	0.1358	0.0550
No log	25.0	175	3.5502	0.82	0.2827	1.1141	0.82	0.7908	0.1460	0.0554
No log	26.0	182	3.5747	0.82	0.2829	1.1727	0.82	0.8027	0.1330	0.0565
No log	27.0	189	3.6091	0.83	0.2787	1.1040	0.83	0.8067	0.1347	0.0563
No log	28.0	196	3.5917	0.82	0.2837	1.1775	0.82	0.7975	0.1513	0.0564
No log	29.0	203	3.6087	0.815	0.2875	1.1448	0.815	0.7998	0.1339	0.0542
No log	30.0	210	3.6018	0.815	0.2819	1.1613	0.815	0.8027	0.1507	0.0535
No log	31.0	217	3.6350	0.815	0.2845	1.2278	0.815	0.7866	0.1401	0.0537
No log	32.0	224	3.6290	0.82	0.2815	1.1528	0.82	0.7950	0.1424	0.0520
No log	33.0	231	3.6642	0.815	0.2865	1.1504	0.815	0.7946	0.1379	0.0542
No log	34.0	238	3.6778	0.815	0.2929	1.2116	0.815	0.7890	0.1437	0.0538
No log	35.0	245	3.6867	0.82	0.2869	1.1547	0.82	0.7904	0.1404	0.0529
No log	36.0	252	3.6931	0.795	0.2946	1.1478	0.795	0.7694	0.1494	0.0543
No log	37.0	259	3.7166	0.82	0.2921	1.2109	0.82	0.7904	0.1489	0.0534
No log	38.0	266	3.7024	0.81	0.2889	1.1516	0.81	0.7888	0.1508	0.0536
No log	39.0	273	3.7353	0.81	0.2943	1.2088	0.81	0.7812	0.1466	0.0537
No log	40.0	280	3.7198	0.82	0.2891	1.1515	0.82	0.8014	0.1285	0.0536
No log	41.0	287	3.7413	0.815	0.2899	1.2124	0.815	0.7959	0.1471	0.0537
No log	42.0	294	3.7272	0.82	0.2896	1.2071	0.82	0.8002	0.1414	0.0532
No log	43.0	301	3.7609	0.815	0.2925	1.2100	0.815	0.7868	0.1486	0.0528
No log	44.0	308	3.7589	0.815	0.2922	1.2074	0.815	0.7877	0.1398	0.0537
No log	45.0	315	3.7820	0.815	0.2961	1.2078	0.815	0.7874	0.1499	0.0535
No log	46.0	322	3.7663	0.82	0.2926	1.2053	0.82	0.8014	0.1369	0.0532
No log	47.0	329	3.7850	0.82	0.2944	1.2079	0.82	0.7904	0.1374	0.0532
No log	48.0	336	3.7802	0.82	0.2935	1.2025	0.82	0.7981	0.1483	0.0537
No log	49.0	343	3.7954	0.82	0.2937	1.2068	0.82	0.7900	0.1354	0.0528
No log	50.0	350	3.7974	0.815	0.2954	1.2020	0.815	0.7907	0.1491	0.0534
No log	51.0	357	3.8081	0.815	0.2965	1.2035	0.815	0.7907	0.1533	0.0533
No log	52.0	364	3.8171	0.815	0.2982	1.2033	0.815	0.7907	0.1466	0.0537
No log	53.0	371	3.8136	0.815	0.2961	1.2035	0.815	0.7907	0.1399	0.0531
No log	54.0	378	3.8244	0.815	0.2977	1.2024	0.815	0.7907	0.1586	0.0538
No log	55.0	385	3.8265	0.815	0.2963	1.2004	0.815	0.7907	0.1506	0.0537
No log	56.0	392	3.8376	0.82	0.2980	1.2011	0.82	0.7964	0.1471	0.0536
No log	57.0	399	3.8428	0.82	0.2982	1.1994	0.82	0.7964	0.1562	0.0535
No log	58.0	406	3.8418	0.82	0.2973	1.2004	0.82	0.7964	0.1484	0.0537
No log	59.0	413	3.8507	0.82	0.2984	1.2009	0.82	0.7931	0.1563	0.0538
No log	60.0	420	3.8560	0.82	0.2989	1.2001	0.82	0.7964	0.1579	0.0540
No log	61.0	427	3.8563	0.82	0.2974	1.1997	0.82	0.7964	0.1560	0.0536
No log	62.0	434	3.8648	0.815	0.2986	1.1995	0.815	0.7907	0.1532	0.0540
No log	63.0	441	3.8682	0.82	0.2991	1.1991	0.82	0.7964	0.1570	0.0536
No log	64.0	448	3.8735	0.82	0.2989	1.1984	0.82	0.7964	0.1481	0.0539
No log	65.0	455	3.8794	0.82	0.3000	1.1981	0.82	0.7964	0.1496	0.0543
No log	66.0	462	3.8824	0.82	0.3002	1.1980	0.82	0.7964	0.1567	0.0539
No log	67.0	469	3.8842	0.82	0.3005	1.1983	0.82	0.7964	0.1438	0.0542
No log	68.0	476	3.8866	0.82	0.3001	1.1978	0.82	0.7964	0.1418	0.0540
No log	69.0	483	3.8912	0.82	0.3003	1.1977	0.82	0.7964	0.1570	0.0541
No log	70.0	490	3.8959	0.82	0.3008	1.1971	0.82	0.7964	0.1445	0.0544
No log	71.0	497	3.8964	0.82	0.3002	1.1977	0.82	0.7964	0.1366	0.0543
3.4649	72.0	504	3.9021	0.82	0.3009	1.1969	0.82	0.7964	0.1471	0.0543
3.4649	73.0	511	3.9052	0.82	0.3015	1.1976	0.82	0.7964	0.1532	0.0546
3.4649	74.0	518	3.9043	0.82	0.3002	1.1973	0.82	0.7964	0.1371	0.0544
3.4649	75.0	525	3.9096	0.82	0.3004	1.1966	0.82	0.7964	0.1417	0.0543
3.4649	76.0	532	3.9099	0.82	0.3010	1.1965	0.82	0.7964	0.1428	0.0545
3.4649	77.0	539	3.9151	0.82	0.3016	1.1963	0.82	0.7964	0.1460	0.0548
3.4649	78.0	546	3.9143	0.82	0.3010	1.1970	0.82	0.7964	0.1447	0.0543
3.4649	79.0	553	3.9164	0.82	0.3014	1.1966	0.82	0.7964	0.1436	0.0545
3.4649	80.0	560	3.9198	0.82	0.3018	1.1965	0.82	0.7964	0.1520	0.0545
3.4649	81.0	567	3.9218	0.82	0.3015	1.1959	0.82	0.7964	0.1440	0.0546
3.4649	82.0	574	3.9236	0.82	0.3018	1.1961	0.82	0.7964	0.1439	0.0546
3.4649	83.0	581	3.9248	0.82	0.3017	1.1959	0.82	0.7964	0.1440	0.0546
3.4649	84.0	588	3.9267	0.82	0.3018	1.1958	0.82	0.7964	0.1442	0.0545
3.4649	85.0	595	3.9286	0.82	0.3019	1.1959	0.82	0.7964	0.1443	0.0546
3.4649	86.0	602	3.9300	0.82	0.3020	1.1958	0.82	0.7964	0.1444	0.0545
3.4649	87.0	609	3.9320	0.82	0.3022	1.1956	0.82	0.7964	0.1446	0.0546
3.4649	88.0	616	3.9327	0.82	0.3022	1.1957	0.82	0.7964	0.1446	0.0545
3.4649	89.0	623	3.9340	0.82	0.3022	1.1955	0.82	0.7964	0.1436	0.0546
3.4649	90.0	630	3.9346	0.82	0.3022	1.1956	0.82	0.7964	0.1447	0.0546
3.4649	91.0	637	3.9360	0.82	0.3023	1.1953	0.82	0.7964	0.1438	0.0546
3.4649	92.0	644	3.9368	0.82	0.3023	1.1954	0.82	0.7964	0.1438	0.0546
3.4649	93.0	651	3.9374	0.82	0.3023	1.1954	0.82	0.7964	0.1437	0.0548
3.4649	94.0	658	3.9380	0.82	0.3023	1.1953	0.82	0.7964	0.1438	0.0548
3.4649	95.0	665	3.9385	0.82	0.3023	1.1953	0.82	0.7964	0.1494	0.0549
3.4649	96.0	672	3.9391	0.82	0.3024	1.1952	0.82	0.7964	0.1494	0.0548
3.4649	97.0	679	3.9393	0.82	0.3024	1.1952	0.82	0.7964	0.1495	0.0548
3.4649	98.0	686	3.9396	0.82	0.3024	1.1952	0.82	0.7964	0.1494	0.0548
3.4649	99.0	693	3.9398	0.82	0.3024	1.1952	0.82	0.7964	0.1494	0.0548
3.4649	100.0	700	3.9399	0.82	0.3024	1.1952	0.82	0.7964	0.1494	0.0548

Framework versions

Transformers 4.26.1
Pytorch 1.13.1.post200
Datasets 2.9.0
Tokenizers 0.13.2

jordyvl/vit-small_tobacco3482_kd_NKD_t1.0_g1.5

vit-small_tobacco3482_kd_NKD_t1.0_g1.5

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions