39-tiny_tobacco3482_kd_NKD_t1.0_g1.5

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: 4.0812
Accuracy: 0.835
Brier Loss: 0.2748
Nll: 1.2215
F1 Micro: 0.835
F1 Macro: 0.8213
Ece: 0.1443
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	5.1938	0.095	1.0201	8.6917	0.095	0.0778	0.3242	0.9009
No log	2.0	14	4.3129	0.13	0.9109	8.2379	0.13	0.0910	0.2544	0.8466
No log	3.0	21	3.9690	0.225	0.8600	6.8547	0.225	0.1401	0.2626	0.6398
No log	4.0	28	3.8651	0.375	0.7978	5.6610	0.375	0.2964	0.3198	0.4692
No log	5.0	35	3.8115	0.465	0.7222	3.4731	0.465	0.3435	0.3007	0.3464
No log	6.0	42	3.7351	0.575	0.6691	2.6672	0.575	0.4736	0.3509	0.2284
No log	7.0	49	3.6913	0.62	0.6152	2.6026	0.62	0.4700	0.3255	0.1827
No log	8.0	56	3.6687	0.68	0.5820	1.9726	0.68	0.5400	0.3735	0.1472
No log	9.0	63	3.6771	0.645	0.5464	1.9938	0.645	0.5211	0.3013	0.1595
No log	10.0	70	3.6759	0.685	0.4884	1.9735	0.685	0.5678	0.2672	0.1278
No log	11.0	77	3.6587	0.71	0.4696	2.0625	0.7100	0.6080	0.2956	0.1115
No log	12.0	84	3.6317	0.72	0.4121	2.2088	0.72	0.6137	0.2372	0.0925
No log	13.0	91	3.6799	0.745	0.4167	2.0639	0.745	0.6372	0.2480	0.0978
No log	14.0	98	3.6191	0.745	0.3850	1.9955	0.745	0.6384	0.2363	0.0728
No log	15.0	105	3.6813	0.715	0.3814	2.0731	0.715	0.6026	0.1995	0.0918
No log	16.0	112	3.6394	0.75	0.3644	1.9093	0.75	0.6492	0.1904	0.0777
No log	17.0	119	3.7661	0.735	0.3786	1.5402	0.735	0.6352	0.2032	0.0982
No log	18.0	126	3.6849	0.79	0.3369	1.8761	0.79	0.6965	0.1954	0.0708
No log	19.0	133	3.6776	0.775	0.3358	1.4981	0.775	0.7021	0.1919	0.0744
No log	20.0	140	3.6814	0.755	0.3546	1.5225	0.755	0.6873	0.1840	0.0794
No log	21.0	147	3.6948	0.775	0.3267	1.4776	0.775	0.7052	0.1630	0.0710
No log	22.0	154	3.7210	0.795	0.3191	1.3634	0.795	0.7383	0.1737	0.0705
No log	23.0	161	3.7231	0.805	0.3062	1.3141	0.805	0.7679	0.1629	0.0665
No log	24.0	168	3.7322	0.815	0.2903	1.2030	0.815	0.7771	0.1789	0.0609
No log	25.0	175	3.7237	0.815	0.3020	1.1721	0.815	0.7947	0.1759	0.0603
No log	26.0	182	3.8243	0.8	0.3138	1.3356	0.8000	0.7699	0.1735	0.0720
No log	27.0	189	3.7675	0.81	0.3038	1.2662	0.81	0.7853	0.1891	0.0699
No log	28.0	196	3.8006	0.81	0.2992	1.3422	0.81	0.7805	0.1709	0.0698
No log	29.0	203	3.7783	0.815	0.3009	1.3322	0.815	0.7959	0.1729	0.0669
No log	30.0	210	3.7547	0.835	0.2775	0.9761	0.835	0.8228	0.1751	0.0566
No log	31.0	217	3.7810	0.82	0.2905	1.1472	0.82	0.7953	0.1670	0.0631
No log	32.0	224	3.7935	0.82	0.2732	1.2016	0.82	0.7967	0.1429	0.0590
No log	33.0	231	3.7871	0.83	0.2774	1.2459	0.83	0.8134	0.1495	0.0562
No log	34.0	238	3.7689	0.815	0.2756	1.1135	0.815	0.7825	0.1609	0.0596
No log	35.0	245	3.8169	0.81	0.2801	1.2621	0.81	0.7880	0.1570	0.0624
No log	36.0	252	3.7973	0.82	0.2729	1.1310	0.82	0.7894	0.1466	0.0585
No log	37.0	259	3.8560	0.835	0.2825	1.3222	0.835	0.8114	0.1466	0.0606
No log	38.0	266	3.8351	0.83	0.2892	1.2548	0.83	0.8178	0.1489	0.0593
No log	39.0	273	3.8258	0.82	0.2711	1.1900	0.82	0.8037	0.1455	0.0589
No log	40.0	280	3.8288	0.815	0.2840	1.2167	0.815	0.7913	0.1574	0.0619
No log	41.0	287	3.8264	0.82	0.2790	1.1737	0.82	0.8020	0.1394	0.0609
No log	42.0	294	3.8276	0.81	0.2797	1.1603	0.81	0.7888	0.1585	0.0580
No log	43.0	301	3.8554	0.815	0.2771	1.1695	0.815	0.7943	0.1310	0.0594
No log	44.0	308	3.8405	0.825	0.2768	1.1593	0.825	0.8149	0.1413	0.0569
No log	45.0	315	3.8640	0.815	0.2891	1.1752	0.815	0.7980	0.1516	0.0590
No log	46.0	322	3.8624	0.825	0.2653	1.1548	0.825	0.8024	0.1384	0.0581
No log	47.0	329	3.8546	0.83	0.2766	1.1634	0.83	0.8106	0.1411	0.0594
No log	48.0	336	3.8652	0.82	0.2805	1.1651	0.82	0.8069	0.1278	0.0581
No log	49.0	343	3.8716	0.83	0.2758	1.1895	0.83	0.8065	0.1486	0.0590
No log	50.0	350	3.8720	0.815	0.2737	1.1709	0.815	0.7937	0.1375	0.0578
No log	51.0	357	3.8812	0.82	0.2762	1.2348	0.82	0.7993	0.1292	0.0600
No log	52.0	364	3.8844	0.805	0.2815	1.0870	0.805	0.7843	0.1525	0.0581
No log	53.0	371	3.8968	0.825	0.2704	1.2235	0.825	0.8011	0.1452	0.0582
No log	54.0	378	3.8996	0.81	0.2788	1.3264	0.81	0.7909	0.1453	0.0573
No log	55.0	385	3.9037	0.81	0.2757	1.2231	0.81	0.7928	0.1307	0.0574
No log	56.0	392	3.9024	0.81	0.2775	1.2369	0.81	0.7869	0.1493	0.0581
No log	57.0	399	3.8951	0.83	0.2722	1.2151	0.83	0.8171	0.1491	0.0556
No log	58.0	406	3.9224	0.82	0.2741	1.2957	0.82	0.8001	0.1351	0.0575
No log	59.0	413	3.9397	0.805	0.2782	1.3017	0.805	0.7870	0.1342	0.0584
No log	60.0	420	3.9250	0.835	0.2721	1.2251	0.835	0.8151	0.1466	0.0570
No log	61.0	427	3.9381	0.825	0.2753	1.2330	0.825	0.8044	0.1384	0.0577
No log	62.0	434	3.9475	0.82	0.2759	1.2171	0.82	0.8054	0.1485	0.0576
No log	63.0	441	3.9591	0.83	0.2761	1.2299	0.83	0.8122	0.1551	0.0568
No log	64.0	448	3.9496	0.835	0.2709	1.2282	0.835	0.8223	0.1397	0.0559
No log	65.0	455	3.9360	0.83	0.2688	1.2238	0.83	0.8171	0.1384	0.0535
No log	66.0	462	3.9594	0.835	0.2733	1.2395	0.835	0.8094	0.1540	0.0563
No log	67.0	469	3.9648	0.84	0.2700	1.2154	0.8400	0.8252	0.1673	0.0557
No log	68.0	476	3.9725	0.83	0.2712	1.2297	0.83	0.8171	0.1248	0.0552
No log	69.0	483	3.9844	0.835	0.2719	1.2243	0.835	0.8151	0.1605	0.0557
No log	70.0	490	3.9845	0.83	0.2699	1.2288	0.83	0.8100	0.1223	0.0553
No log	71.0	497	3.9986	0.835	0.2729	1.2206	0.835	0.8223	0.1381	0.0556
3.4116	72.0	504	3.9973	0.835	0.2727	1.2242	0.835	0.8223	0.1446	0.0553
3.4116	73.0	511	4.0092	0.835	0.2733	1.2226	0.835	0.8223	0.1482	0.0554
3.4116	74.0	518	4.0072	0.83	0.2714	1.2248	0.83	0.8152	0.1219	0.0549
3.4116	75.0	525	4.0168	0.835	0.2742	1.2200	0.835	0.8223	0.1329	0.0551
3.4116	76.0	532	4.0223	0.835	0.2737	1.2248	0.835	0.8213	0.1380	0.0552
3.4116	77.0	539	4.0250	0.84	0.2719	1.2208	0.8400	0.8252	0.1405	0.0551
3.4116	78.0	546	4.0338	0.835	0.2745	1.2242	0.835	0.8213	0.1536	0.0551
3.4116	79.0	553	4.0380	0.835	0.2740	1.2234	0.835	0.8213	0.1494	0.0552
3.4116	80.0	560	4.0445	0.835	0.2744	1.2223	0.835	0.8213	0.1500	0.0555
3.4116	81.0	567	4.0449	0.835	0.2735	1.2209	0.835	0.8213	0.1504	0.0552
3.4116	82.0	574	4.0515	0.835	0.2747	1.2228	0.835	0.8213	0.1526	0.0549
3.4116	83.0	581	4.0534	0.835	0.2743	1.2226	0.835	0.8213	0.1501	0.0548
3.4116	84.0	588	4.0572	0.835	0.2740	1.2225	0.835	0.8213	0.1447	0.0550
3.4116	85.0	595	4.0605	0.835	0.2743	1.2222	0.835	0.8213	0.1466	0.0548
3.4116	86.0	602	4.0621	0.835	0.2744	1.2215	0.835	0.8213	0.1427	0.0548
3.4116	87.0	609	4.0653	0.835	0.2745	1.2214	0.835	0.8213	0.1439	0.0549
3.4116	88.0	616	4.0673	0.835	0.2746	1.2217	0.835	0.8213	0.1410	0.0548
3.4116	89.0	623	4.0705	0.835	0.2748	1.2214	0.835	0.8213	0.1440	0.0549
3.4116	90.0	630	4.0717	0.835	0.2744	1.2217	0.835	0.8213	0.1426	0.0547
3.4116	91.0	637	4.0740	0.835	0.2747	1.2217	0.835	0.8213	0.1432	0.0548
3.4116	92.0	644	4.0753	0.835	0.2748	1.2217	0.835	0.8213	0.1442	0.0547
3.4116	93.0	651	4.0763	0.835	0.2746	1.2214	0.835	0.8213	0.1434	0.0546
3.4116	94.0	658	4.0777	0.835	0.2746	1.2213	0.835	0.8213	0.1433	0.0547
3.4116	95.0	665	4.0788	0.835	0.2747	1.2217	0.835	0.8213	0.1442	0.0547
3.4116	96.0	672	4.0800	0.835	0.2748	1.2217	0.835	0.8213	0.1466	0.0547
3.4116	97.0	679	4.0802	0.835	0.2747	1.2215	0.835	0.8213	0.1435	0.0547
3.4116	98.0	686	4.0808	0.835	0.2747	1.2214	0.835	0.8213	0.1435	0.0547
3.4116	99.0	693	4.0811	0.835	0.2748	1.2214	0.835	0.8213	0.1443	0.0547
3.4116	100.0	700	4.0812	0.835	0.2748	1.2215	0.835	0.8213	0.1443	0.0548

Framework versions

Transformers 4.26.1
Pytorch 1.13.1.post200
Datasets 2.9.0
Tokenizers 0.13.2

39-tiny_tobacco3482_kd_NKD_t1.0_g1.5

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

NSDT 3DConvert

UnrealSynth

DreamTexture.js