vit-small_tobacco3482_kd_CE

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: 0.7692
Accuracy: 0.845
Brier Loss: 0.2469
Nll: 1.1078
F1 Micro: 0.845
F1 Macro: 0.8517
Ece: 0.1239
Aurc: 0.0373

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	2.2384	0.215	0.8750	5.2607	0.2150	0.1263	0.2515	0.6887
No log	2.0	14	1.8043	0.395	0.7402	3.6938	0.395	0.2230	0.2859	0.4088
No log	3.0	21	1.2581	0.61	0.5613	2.0209	0.61	0.5557	0.2634	0.2084
No log	4.0	28	0.8820	0.7	0.4017	1.7039	0.7	0.6574	0.2397	0.1123
No log	5.0	35	0.8338	0.74	0.3807	1.7427	0.74	0.7425	0.2155	0.1001
No log	6.0	42	0.7026	0.775	0.3202	1.4789	0.775	0.7666	0.1781	0.0771
No log	7.0	49	0.7935	0.77	0.3635	1.5766	0.7700	0.7840	0.2029	0.0857
No log	8.0	56	0.6819	0.8	0.3047	1.3800	0.8000	0.7987	0.1758	0.0738
No log	9.0	63	0.7826	0.775	0.3434	1.5345	0.775	0.7888	0.1910	0.0807
No log	10.0	70	0.8752	0.775	0.3392	1.5110	0.775	0.7813	0.1731	0.0737
No log	11.0	77	1.0440	0.72	0.4285	1.7284	0.72	0.7241	0.2416	0.1094
No log	12.0	84	0.8109	0.785	0.3411	1.3933	0.785	0.7806	0.1694	0.0690
No log	13.0	91	0.9980	0.76	0.3852	1.5143	0.76	0.6919	0.1917	0.0892
No log	14.0	98	1.0056	0.775	0.3773	1.4735	0.775	0.7733	0.2060	0.0871
No log	15.0	105	1.2081	0.75	0.4260	1.5653	0.75	0.7373	0.2110	0.0983
No log	16.0	112	1.1463	0.78	0.3781	1.5472	0.78	0.7906	0.1736	0.1004
No log	17.0	119	0.9384	0.825	0.2973	1.5527	0.825	0.8334	0.1529	0.0648
No log	18.0	126	0.9258	0.785	0.3464	1.2875	0.785	0.7832	0.1694	0.0562
No log	19.0	133	1.1667	0.8	0.3406	1.8919	0.8000	0.8038	0.1518	0.0705
No log	20.0	140	0.9351	0.81	0.3116	1.4283	0.81	0.8084	0.1532	0.0628
No log	21.0	147	1.2016	0.77	0.4040	1.2958	0.7700	0.7606	0.2051	0.0827
No log	22.0	154	1.3592	0.765	0.4040	1.6059	0.765	0.7645	0.1941	0.1163
No log	23.0	161	0.9921	0.805	0.3374	1.6304	0.805	0.8029	0.1710	0.0558
No log	24.0	168	0.8805	0.83	0.2953	1.1996	0.83	0.8189	0.1547	0.0611
No log	25.0	175	0.9926	0.815	0.3148	1.2949	0.815	0.8050	0.1638	0.0606
No log	26.0	182	1.0838	0.83	0.3171	1.3327	0.83	0.8265	0.1632	0.0732
No log	27.0	189	1.1845	0.8	0.3382	1.3456	0.8000	0.7942	0.1814	0.0798
No log	28.0	196	0.9800	0.83	0.2999	1.3172	0.83	0.8275	0.1563	0.0798
No log	29.0	203	0.9653	0.85	0.2724	1.3303	0.85	0.8531	0.1415	0.0556
No log	30.0	210	0.9896	0.85	0.2837	1.3282	0.85	0.8494	0.1373	0.0596
No log	31.0	217	0.9196	0.84	0.2844	1.2157	0.8400	0.8437	0.1516	0.0508
No log	32.0	224	0.9701	0.83	0.3062	1.2264	0.83	0.8364	0.1608	0.0554
No log	33.0	231	0.7464	0.865	0.2353	1.1321	0.865	0.8613	0.1265	0.0432
No log	34.0	238	0.7593	0.865	0.2367	1.1160	0.865	0.8649	0.1322	0.0430
No log	35.0	245	0.7450	0.855	0.2465	1.0615	0.855	0.8536	0.1279	0.0413
No log	36.0	252	0.7389	0.845	0.2546	1.0563	0.845	0.8429	0.1266	0.0417
No log	37.0	259	0.7332	0.845	0.2542	1.0549	0.845	0.8452	0.1293	0.0413
No log	38.0	266	0.7328	0.85	0.2531	1.0554	0.85	0.8490	0.1331	0.0407
No log	39.0	273	0.7342	0.85	0.2514	1.0558	0.85	0.8490	0.1339	0.0398
No log	40.0	280	0.7367	0.855	0.2498	1.0564	0.855	0.8529	0.1362	0.0391
No log	41.0	287	0.7395	0.855	0.2489	1.0574	0.855	0.8529	0.1307	0.0392
No log	42.0	294	0.7412	0.855	0.2480	1.0598	0.855	0.8529	0.1237	0.0393
No log	43.0	301	0.7434	0.855	0.2475	1.0635	0.855	0.8550	0.1161	0.0392
No log	44.0	308	0.7453	0.855	0.2473	1.0725	0.855	0.8550	0.1237	0.0392
No log	45.0	315	0.7462	0.855	0.2471	1.1225	0.855	0.8550	0.1205	0.0391
No log	46.0	322	0.7471	0.855	0.2468	1.1219	0.855	0.8550	0.1155	0.0391
No log	47.0	329	0.7481	0.85	0.2466	1.1213	0.85	0.8519	0.1283	0.0390
No log	48.0	336	0.7492	0.85	0.2464	1.1207	0.85	0.8519	0.1334	0.0388
No log	49.0	343	0.7504	0.85	0.2464	1.1203	0.85	0.8519	0.1379	0.0387
No log	50.0	350	0.7515	0.85	0.2465	1.1201	0.85	0.8519	0.1267	0.0387
No log	51.0	357	0.7523	0.85	0.2464	1.1198	0.85	0.8519	0.1265	0.0385
No log	52.0	364	0.7532	0.85	0.2463	1.1194	0.85	0.8519	0.1201	0.0385
No log	53.0	371	0.7534	0.855	0.2461	1.1189	0.855	0.8602	0.1266	0.0384
No log	54.0	378	0.7542	0.855	0.2460	1.1185	0.855	0.8602	0.1279	0.0386
No log	55.0	385	0.7547	0.855	0.2459	1.1180	0.855	0.8602	0.1332	0.0381
No log	56.0	392	0.7556	0.855	0.2460	1.1176	0.855	0.8602	0.1256	0.0380
No log	57.0	399	0.7564	0.855	0.2460	1.1171	0.855	0.8602	0.1252	0.0381
No log	58.0	406	0.7571	0.855	0.2461	1.1166	0.855	0.8602	0.1231	0.0379
No log	59.0	413	0.7581	0.855	0.2463	1.1162	0.855	0.8602	0.1295	0.0378
No log	60.0	420	0.7588	0.855	0.2464	1.1159	0.855	0.8602	0.1224	0.0378
No log	61.0	427	0.7594	0.855	0.2465	1.1155	0.855	0.8602	0.1226	0.0378
No log	62.0	434	0.7598	0.855	0.2464	1.1152	0.855	0.8602	0.1231	0.0378
No log	63.0	441	0.7605	0.855	0.2465	1.1149	0.855	0.8602	0.1231	0.0378
No log	64.0	448	0.7610	0.855	0.2465	1.1144	0.855	0.8602	0.1222	0.0377
No log	65.0	455	0.7618	0.855	0.2466	1.1140	0.855	0.8602	0.1229	0.0377
No log	66.0	462	0.7625	0.855	0.2468	1.1137	0.855	0.8602	0.1317	0.0378
No log	67.0	469	0.7630	0.855	0.2468	1.1133	0.855	0.8602	0.1317	0.0377
No log	68.0	476	0.7633	0.855	0.2468	1.1130	0.855	0.8602	0.1317	0.0377
No log	69.0	483	0.7636	0.855	0.2468	1.1127	0.855	0.8602	0.1318	0.0376
No log	70.0	490	0.7640	0.855	0.2468	1.1124	0.855	0.8602	0.1319	0.0376
No log	71.0	497	0.7645	0.85	0.2468	1.1120	0.85	0.8548	0.1280	0.0375
0.1221	72.0	504	0.7649	0.85	0.2468	1.1116	0.85	0.8548	0.1293	0.0375
0.1221	73.0	511	0.7653	0.85	0.2469	1.1113	0.85	0.8548	0.1293	0.0374
0.1221	74.0	518	0.7656	0.85	0.2469	1.1111	0.85	0.8548	0.1293	0.0373
0.1221	75.0	525	0.7659	0.85	0.2469	1.1108	0.85	0.8548	0.1208	0.0373
0.1221	76.0	532	0.7662	0.85	0.2469	1.1106	0.85	0.8548	0.1207	0.0374
0.1221	77.0	539	0.7664	0.85	0.2469	1.1104	0.85	0.8548	0.1207	0.0374
0.1221	78.0	546	0.7665	0.85	0.2469	1.1102	0.85	0.8548	0.1301	0.0375
0.1221	79.0	553	0.7667	0.85	0.2469	1.1100	0.85	0.8548	0.1301	0.0375
0.1221	80.0	560	0.7668	0.85	0.2468	1.1097	0.85	0.8548	0.1301	0.0374
0.1221	81.0	567	0.7669	0.85	0.2468	1.1095	0.85	0.8548	0.1300	0.0374
0.1221	82.0	574	0.7672	0.85	0.2468	1.1094	0.85	0.8548	0.1301	0.0374
0.1221	83.0	581	0.7674	0.85	0.2469	1.1092	0.85	0.8548	0.1301	0.0374
0.1221	84.0	588	0.7678	0.85	0.2469	1.1090	0.85	0.8548	0.1302	0.0374
0.1221	85.0	595	0.7678	0.85	0.2468	1.1089	0.85	0.8548	0.1284	0.0373
0.1221	86.0	602	0.7679	0.845	0.2468	1.1087	0.845	0.8517	0.1243	0.0373
0.1221	87.0	609	0.7681	0.845	0.2468	1.1086	0.845	0.8517	0.1244	0.0373
0.1221	88.0	616	0.7683	0.845	0.2468	1.1084	0.845	0.8517	0.1244	0.0373
0.1221	89.0	623	0.7685	0.845	0.2468	1.1083	0.845	0.8517	0.1243	0.0373
0.1221	90.0	630	0.7687	0.845	0.2469	1.1083	0.845	0.8517	0.1239	0.0372
0.1221	91.0	637	0.7688	0.845	0.2469	1.1082	0.845	0.8517	0.1239	0.0372
0.1221	92.0	644	0.7689	0.845	0.2469	1.1082	0.845	0.8517	0.1239	0.0373
0.1221	93.0	651	0.7690	0.845	0.2469	1.1081	0.845	0.8517	0.1239	0.0373
0.1221	94.0	658	0.7690	0.845	0.2469	1.1080	0.845	0.8517	0.1239	0.0373
0.1221	95.0	665	0.7691	0.845	0.2469	1.1080	0.845	0.8517	0.1239	0.0373
0.1221	96.0	672	0.7692	0.845	0.2469	1.1079	0.845	0.8517	0.1239	0.0373
0.1221	97.0	679	0.7692	0.845	0.2469	1.1079	0.845	0.8517	0.1239	0.0373
0.1221	98.0	686	0.7692	0.845	0.2469	1.1078	0.845	0.8517	0.1239	0.0373
0.1221	99.0	693	0.7692	0.845	0.2469	1.1078	0.845	0.8517	0.1239	0.0373
0.1221	100.0	700	0.7692	0.845	0.2469	1.1078	0.845	0.8517	0.1239	0.0373

Framework versions

Transformers 4.26.1
Pytorch 1.13.1.post200
Datasets 2.9.0
Tokenizers 0.13.2

jordyvl/vit-small_tobacco3482_kd_CE

vit-small_tobacco3482_kd_CE

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions