171-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: 4.3051
Accuracy: 0.815
Brier Loss: 0.3074
Nll: 1.7785
F1 Micro: 0.815
F1 Macro: 0.8049
Ece: 0.1516
Aurc: 0.0489

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: 64
eval_batch_size: 64
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	13	5.0451	0.22	0.8885	8.4566	0.22	0.1247	0.2851	0.7928
No log	2.0	26	4.5055	0.385	0.7765	3.9655	0.3850	0.3066	0.3049	0.4178
No log	3.0	39	4.3096	0.52	0.6691	3.8146	0.52	0.3950	0.3177	0.2860
No log	4.0	52	4.1755	0.575	0.5907	2.9444	0.575	0.4546	0.2729	0.2029
No log	5.0	65	4.0437	0.675	0.5104	2.4241	0.675	0.5995	0.2991	0.1354
No log	6.0	78	4.0642	0.69	0.4602	2.3471	0.69	0.5925	0.2798	0.1256
No log	7.0	91	4.0104	0.695	0.4319	2.2902	0.695	0.6109	0.2430	0.1101
No log	8.0	104	4.1702	0.7	0.4296	2.5778	0.7	0.6065	0.2231	0.1201
No log	9.0	117	4.2785	0.695	0.4433	2.7331	0.695	0.6269	0.2296	0.1283
No log	10.0	130	3.9853	0.725	0.3705	2.0880	0.7250	0.6477	0.1971	0.0874
No log	11.0	143	3.9595	0.725	0.3506	2.1144	0.7250	0.6431	0.1650	0.0750
No log	12.0	156	3.8678	0.735	0.3504	2.0683	0.735	0.6839	0.2047	0.0764
No log	13.0	169	3.9641	0.745	0.3520	2.0788	0.745	0.6754	0.1899	0.0837
No log	14.0	182	4.0188	0.725	0.3639	2.3771	0.7250	0.6643	0.1740	0.0893
No log	15.0	195	3.8558	0.765	0.3342	1.4620	0.765	0.7097	0.1866	0.0696
No log	16.0	208	3.9103	0.79	0.3416	1.7139	0.79	0.7662	0.2043	0.0770
No log	17.0	221	4.0320	0.795	0.3548	1.8525	0.795	0.7690	0.1901	0.0924
No log	18.0	234	3.8974	0.79	0.3264	1.8646	0.79	0.7582	0.1656	0.0739
No log	19.0	247	3.8235	0.815	0.3074	1.4771	0.815	0.8185	0.1825	0.0617
No log	20.0	260	3.8918	0.805	0.3150	1.6824	0.805	0.7893	0.1859	0.0631
No log	21.0	273	3.8919	0.785	0.3161	1.7951	0.785	0.7725	0.1450	0.0701
No log	22.0	286	3.8626	0.795	0.3121	1.6707	0.795	0.7832	0.1570	0.0684
No log	23.0	299	3.8132	0.825	0.2906	1.4511	0.825	0.8097	0.1552	0.0564
No log	24.0	312	3.8680	0.81	0.3048	1.9348	0.81	0.8027	0.1572	0.0611
No log	25.0	325	3.8305	0.81	0.2954	1.5734	0.81	0.7999	0.1645	0.0556
No log	26.0	338	3.8050	0.81	0.2965	1.7904	0.81	0.8013	0.1495	0.0546
No log	27.0	351	3.9524	0.79	0.3212	2.0459	0.79	0.7846	0.1643	0.0669
No log	28.0	364	3.9299	0.81	0.3076	1.7819	0.81	0.7967	0.1393	0.0601
No log	29.0	377	3.9315	0.805	0.3158	2.0697	0.805	0.8046	0.1618	0.0663
No log	30.0	390	3.8141	0.825	0.2853	1.9079	0.825	0.8150	0.1487	0.0528
No log	31.0	403	3.8682	0.815	0.2932	1.9092	0.815	0.8030	0.1448	0.0585
No log	32.0	416	3.8275	0.82	0.2823	1.6793	0.82	0.8043	0.1459	0.0508
No log	33.0	429	3.8782	0.82	0.2895	1.6565	0.82	0.8077	0.1465	0.0542
No log	34.0	442	3.8433	0.825	0.2891	1.6481	0.825	0.8157	0.1467	0.0525
No log	35.0	455	3.8403	0.82	0.2891	1.5960	0.82	0.8090	0.1398	0.0497
No log	36.0	468	3.8627	0.81	0.2848	1.6935	0.81	0.8015	0.1557	0.0471
No log	37.0	481	3.8992	0.81	0.2937	1.8237	0.81	0.7991	0.1511	0.0515
No log	38.0	494	3.9662	0.82	0.2978	1.8392	0.82	0.8143	0.1503	0.0527
3.5354	39.0	507	3.9440	0.825	0.2899	1.7818	0.825	0.8159	0.1454	0.0540
3.5354	40.0	520	3.9479	0.81	0.2959	1.7465	0.81	0.7986	0.1504	0.0501
3.5354	41.0	533	3.9760	0.815	0.2964	1.7821	0.815	0.8049	0.1519	0.0522
3.5354	42.0	546	3.9696	0.82	0.2906	1.7671	0.82	0.8127	0.1468	0.0503
3.5354	43.0	559	4.0107	0.81	0.2994	1.8207	0.81	0.7986	0.1474	0.0517
3.5354	44.0	572	3.9970	0.815	0.2913	1.7706	0.815	0.8049	0.1465	0.0504
3.5354	45.0	585	3.9890	0.815	0.2886	1.6384	0.815	0.8049	0.1516	0.0495
3.5354	46.0	598	4.0585	0.82	0.3006	1.7773	0.82	0.8127	0.1522	0.0518
3.5354	47.0	611	4.0448	0.825	0.2925	1.8226	0.825	0.8109	0.1540	0.0505
3.5354	48.0	624	4.0918	0.815	0.3016	1.8403	0.815	0.8049	0.1492	0.0512
3.5354	49.0	637	4.0677	0.82	0.2971	1.8256	0.82	0.8127	0.1396	0.0493
3.5354	50.0	650	4.0831	0.815	0.2986	1.8232	0.815	0.8049	0.1479	0.0513
3.5354	51.0	663	4.0846	0.815	0.2994	1.8268	0.815	0.8049	0.1525	0.0496
3.5354	52.0	676	4.0828	0.82	0.2978	1.7538	0.82	0.8127	0.1425	0.0486
3.5354	53.0	689	4.0890	0.815	0.3004	1.7552	0.815	0.8049	0.1491	0.0485
3.5354	54.0	702	4.1299	0.815	0.3029	1.8902	0.815	0.8049	0.1614	0.0506
3.5354	55.0	715	4.1200	0.815	0.3016	1.8279	0.815	0.8049	0.1510	0.0499
3.5354	56.0	728	4.1196	0.815	0.3008	1.8883	0.815	0.8049	0.1503	0.0503
3.5354	57.0	741	4.1200	0.815	0.3003	1.7620	0.815	0.8049	0.1499	0.0490
3.5354	58.0	754	4.1419	0.815	0.3017	1.8463	0.815	0.8049	0.1459	0.0499
3.5354	59.0	767	4.1527	0.815	0.3041	1.8269	0.815	0.8049	0.1618	0.0496
3.5354	60.0	780	4.1362	0.815	0.3002	1.7666	0.815	0.8049	0.1461	0.0489
3.5354	61.0	793	4.1470	0.815	0.3009	1.8213	0.815	0.8049	0.1471	0.0491
3.5354	62.0	806	4.1503	0.815	0.2991	1.8235	0.815	0.8049	0.1604	0.0496
3.5354	63.0	819	4.1544	0.815	0.3003	1.7546	0.815	0.8049	0.1518	0.0487
3.5354	64.0	832	4.1713	0.815	0.3023	1.8223	0.815	0.8049	0.1543	0.0499
3.5354	65.0	845	4.1716	0.815	0.3010	1.8213	0.815	0.8049	0.1485	0.0494
3.5354	66.0	858	4.1956	0.815	0.3042	1.8287	0.815	0.8049	0.1637	0.0496
3.5354	67.0	871	4.1845	0.815	0.3018	1.8259	0.815	0.8049	0.1519	0.0488
3.5354	68.0	884	4.2055	0.815	0.3037	1.8339	0.815	0.8049	0.1504	0.0496
3.5354	69.0	897	4.2079	0.815	0.3039	1.8281	0.815	0.8049	0.1554	0.0491
3.5354	70.0	910	4.2125	0.815	0.3034	1.7637	0.815	0.8049	0.1500	0.0490
3.5354	71.0	923	4.2179	0.815	0.3035	1.8254	0.815	0.8049	0.1531	0.0492
3.5354	72.0	936	4.2270	0.815	0.3040	1.8270	0.815	0.8049	0.1528	0.0493
3.5354	73.0	949	4.2294	0.815	0.3041	1.8260	0.815	0.8049	0.1531	0.0488
3.5354	74.0	962	4.2383	0.815	0.3043	1.8261	0.815	0.8049	0.1513	0.0492
3.5354	75.0	975	4.2441	0.815	0.3051	1.7691	0.815	0.8049	0.1539	0.0488
3.5354	76.0	988	4.2500	0.815	0.3051	1.8287	0.815	0.8049	0.1540	0.0490
3.192	77.0	1001	4.2538	0.815	0.3053	1.8273	0.815	0.8049	0.1542	0.0490
3.192	78.0	1014	4.2573	0.815	0.3055	1.8281	0.815	0.8049	0.1541	0.0491
3.192	79.0	1027	4.2603	0.815	0.3054	1.8275	0.815	0.8049	0.1544	0.0490
3.192	80.0	1040	4.2673	0.815	0.3060	1.8277	0.815	0.8049	0.1544	0.0489
3.192	81.0	1053	4.2697	0.815	0.3060	1.8272	0.815	0.8049	0.1500	0.0489
3.192	82.0	1066	4.2747	0.815	0.3064	1.7765	0.815	0.8049	0.1544	0.0489
3.192	83.0	1079	4.2769	0.815	0.3063	1.8273	0.815	0.8049	0.1503	0.0489
3.192	84.0	1092	4.2824	0.815	0.3066	1.8278	0.815	0.8049	0.1548	0.0491
3.192	85.0	1105	4.2842	0.815	0.3066	1.8276	0.815	0.8049	0.1506	0.0489
3.192	86.0	1118	4.2883	0.815	0.3070	1.8281	0.815	0.8049	0.1508	0.0488
3.192	87.0	1131	4.2907	0.815	0.3071	1.7730	0.815	0.8049	0.1548	0.0489
3.192	88.0	1144	4.2919	0.815	0.3070	1.7739	0.815	0.8049	0.1513	0.0489
3.192	89.0	1157	4.2943	0.815	0.3071	1.8281	0.815	0.8049	0.1514	0.0489
3.192	90.0	1170	4.2954	0.815	0.3070	1.8280	0.815	0.8049	0.1508	0.0489
3.192	91.0	1183	4.2976	0.815	0.3071	1.8282	0.815	0.8049	0.1514	0.0489
3.192	92.0	1196	4.2985	0.815	0.3070	1.7799	0.815	0.8049	0.1509	0.0489
3.192	93.0	1209	4.3000	0.815	0.3072	1.7832	0.815	0.8049	0.1514	0.0489
3.192	94.0	1222	4.3016	0.815	0.3073	1.7775	0.815	0.8049	0.1516	0.0489
3.192	95.0	1235	4.3025	0.815	0.3072	1.8282	0.815	0.8049	0.1510	0.0489
3.192	96.0	1248	4.3030	0.815	0.3073	1.7778	0.815	0.8049	0.1510	0.0489
3.192	97.0	1261	4.3042	0.815	0.3073	1.7770	0.815	0.8049	0.1516	0.0489
3.192	98.0	1274	4.3047	0.815	0.3074	1.7826	0.815	0.8049	0.1516	0.0489
3.192	99.0	1287	4.3051	0.815	0.3074	1.7777	0.815	0.8049	0.1516	0.0489
3.192	100.0	1300	4.3051	0.815	0.3074	1.7785	0.815	0.8049	0.1516	0.0489

Framework versions

Transformers 4.26.1
Pytorch 1.13.1.post200
Datasets 2.9.0
Tokenizers 0.13.2

jordyvl/171-tiny_tobacco3482_kd_NKD_t1.0_g1.5

171-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