60-tiny_tobacco3482_hint_

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: 63.5396
Accuracy: 0.84
Brier Loss: 0.3043
Nll: 1.1495
F1 Micro: 0.8400
F1 Macro: 0.8244
Ece: 0.1568
Aurc: 0.0457

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	66.6267	0.26	0.8706	4.9000	0.26	0.1920	0.2904	0.7812
No log	2.0	50	65.8408	0.54	0.5951	2.8543	0.54	0.4535	0.2542	0.2567
No log	3.0	75	65.3708	0.675	0.4400	1.6094	0.675	0.6134	0.2395	0.1333
No log	4.0	100	65.0889	0.76	0.3809	1.5505	0.76	0.7422	0.2333	0.1125
No log	5.0	125	64.7800	0.8	0.3080	1.7523	0.8000	0.7663	0.1734	0.0708
No log	6.0	150	64.6296	0.78	0.3286	1.7771	0.78	0.7427	0.1752	0.0642
No log	7.0	175	64.3879	0.765	0.3584	1.7400	0.765	0.6986	0.1799	0.0937
No log	8.0	200	64.4361	0.72	0.4640	1.4368	0.72	0.7385	0.2350	0.1314
No log	9.0	225	64.2223	0.76	0.3846	1.6420	0.76	0.7417	0.2006	0.0915
No log	10.0	250	64.2618	0.725	0.4268	1.6667	0.7250	0.7132	0.2131	0.1110
No log	11.0	275	64.2839	0.7	0.4830	1.7975	0.7	0.6829	0.2213	0.1207
No log	12.0	300	64.0218	0.785	0.3523	1.7098	0.785	0.7363	0.1742	0.0702
No log	13.0	325	63.8071	0.78	0.3218	1.4587	0.78	0.7574	0.1640	0.0674
No log	14.0	350	64.7387	0.645	0.5871	2.0188	0.645	0.6360	0.2996	0.1765
No log	15.0	375	64.2173	0.765	0.3832	1.8093	0.765	0.6909	0.1978	0.0892
No log	16.0	400	64.2233	0.765	0.3897	1.4456	0.765	0.7432	0.1983	0.0805
No log	17.0	425	63.7977	0.825	0.2971	1.4248	0.825	0.8057	0.1583	0.0546
No log	18.0	450	63.5818	0.82	0.2983	1.3079	0.82	0.7936	0.1532	0.0474
No log	19.0	475	64.1935	0.78	0.3764	1.6662	0.78	0.7618	0.1911	0.0669
63.0313	20.0	500	63.6054	0.825	0.2871	1.4054	0.825	0.8118	0.1605	0.0520
63.0313	21.0	525	63.6316	0.79	0.3258	1.3131	0.79	0.7714	0.1632	0.0485
63.0313	22.0	550	63.6978	0.84	0.2935	1.2425	0.8400	0.8236	0.1508	0.0586
63.0313	23.0	575	63.8266	0.825	0.3117	1.5766	0.825	0.8019	0.1550	0.0554
63.0313	24.0	600	63.6750	0.825	0.3130	1.1848	0.825	0.8158	0.1553	0.0462
63.0313	25.0	625	63.8469	0.82	0.3259	1.3997	0.82	0.8007	0.1603	0.0564
63.0313	26.0	650	63.7656	0.815	0.3285	1.2752	0.815	0.7969	0.1656	0.0535
63.0313	27.0	675	63.8074	0.805	0.3455	1.1282	0.805	0.7870	0.1732	0.0542
63.0313	28.0	700	63.8411	0.81	0.3437	1.1501	0.81	0.7917	0.1759	0.0529
63.0313	29.0	725	63.8158	0.81	0.3345	1.1519	0.81	0.7901	0.1706	0.0544
63.0313	30.0	750	63.7917	0.815	0.3383	1.2013	0.815	0.8006	0.1706	0.0557
63.0313	31.0	775	63.7855	0.815	0.3396	1.2088	0.815	0.7974	0.1687	0.0551
63.0313	32.0	800	63.8003	0.825	0.3297	1.2233	0.825	0.8091	0.1694	0.0547
63.0313	33.0	825	63.8029	0.815	0.3405	1.2628	0.815	0.8007	0.1729	0.0547
63.0313	34.0	850	63.7752	0.81	0.3352	1.2587	0.81	0.7979	0.1727	0.0574
63.0313	35.0	875	63.7800	0.815	0.3346	1.1948	0.815	0.7977	0.1679	0.0560
63.0313	36.0	900	63.7885	0.825	0.3313	1.2728	0.825	0.8173	0.1591	0.0569
63.0313	37.0	925	63.7730	0.815	0.3354	1.2726	0.815	0.8027	0.1689	0.0555
63.0313	38.0	950	63.8327	0.815	0.3405	1.4350	0.815	0.8043	0.1675	0.0632
63.0313	39.0	975	63.7324	0.785	0.3686	1.6439	0.785	0.7745	0.1808	0.0666
61.6786	40.0	1000	63.8625	0.765	0.3946	1.6127	0.765	0.7727	0.1961	0.0723
61.6786	41.0	1025	64.1254	0.765	0.3904	1.5456	0.765	0.7570	0.2020	0.0850
61.6786	42.0	1050	63.6201	0.78	0.3728	1.4198	0.78	0.7447	0.1869	0.0647
61.6786	43.0	1075	63.6033	0.835	0.2968	1.5430	0.835	0.8059	0.1574	0.0479
61.6786	44.0	1100	63.6777	0.795	0.3606	1.3542	0.795	0.7638	0.1806	0.0529
61.6786	45.0	1125	63.5747	0.83	0.2996	1.5403	0.83	0.8079	0.1450	0.0504
61.6786	46.0	1150	63.6022	0.805	0.3389	1.3842	0.805	0.7791	0.1794	0.0466
61.6786	47.0	1175	63.6342	0.81	0.3346	1.2861	0.81	0.7811	0.1678	0.0476
61.6786	48.0	1200	63.6065	0.81	0.3298	1.2911	0.81	0.7807	0.1654	0.0465
61.6786	49.0	1225	63.5937	0.815	0.3260	1.3576	0.815	0.7844	0.1613	0.0467
61.6786	50.0	1250	63.6029	0.815	0.3241	1.2826	0.815	0.7844	0.1662	0.0467
61.6786	51.0	1275	63.5947	0.81	0.3232	1.4156	0.81	0.7789	0.1631	0.0471
61.6786	52.0	1300	63.6501	0.81	0.3268	1.4148	0.81	0.7785	0.1703	0.0468
61.6786	53.0	1325	63.6207	0.81	0.3207	1.2785	0.81	0.7785	0.1698	0.0479
61.6786	54.0	1350	63.6021	0.815	0.3233	1.3519	0.815	0.7818	0.1629	0.0456
61.6786	55.0	1375	63.6128	0.815	0.3207	1.2837	0.815	0.7818	0.1641	0.0474
61.6786	56.0	1400	63.5974	0.81	0.3194	1.3542	0.81	0.7789	0.1679	0.0474
61.6786	57.0	1425	63.6173	0.81	0.3260	1.2907	0.81	0.7761	0.1653	0.0486
61.6786	58.0	1450	63.6057	0.81	0.3163	1.2981	0.81	0.7789	0.1651	0.0471
61.6786	59.0	1475	63.6052	0.81	0.3197	1.3444	0.81	0.7789	0.1680	0.0467
61.52	60.0	1500	63.5865	0.82	0.3143	1.2748	0.82	0.7920	0.1617	0.0465
61.52	61.0	1525	63.5754	0.82	0.3126	1.2677	0.82	0.7920	0.1595	0.0468
61.52	62.0	1550	63.5876	0.815	0.3120	1.2691	0.815	0.7879	0.1567	0.0478
61.52	63.0	1575	63.6040	0.82	0.3110	1.2632	0.82	0.7920	0.1526	0.0472
61.52	64.0	1600	63.5956	0.82	0.3111	1.1976	0.82	0.7963	0.1592	0.0468
61.52	65.0	1625	63.5792	0.815	0.3095	1.1928	0.815	0.7879	0.1571	0.0469
61.52	66.0	1650	63.5704	0.82	0.3086	1.2509	0.82	0.7936	0.1543	0.0467
61.52	67.0	1675	63.5918	0.82	0.3118	1.2536	0.82	0.7936	0.1619	0.0471
61.52	68.0	1700	63.5741	0.82	0.3072	1.2491	0.82	0.7963	0.1562	0.0465
61.52	69.0	1725	63.5581	0.825	0.3085	1.2490	0.825	0.8021	0.1566	0.0460
61.52	70.0	1750	63.5796	0.82	0.3087	1.2456	0.82	0.7963	0.1556	0.0471
61.52	71.0	1775	63.5776	0.825	0.3073	1.2530	0.825	0.8021	0.1571	0.0474
61.52	72.0	1800	63.5524	0.825	0.3064	1.2402	0.825	0.8021	0.1555	0.0465
61.52	73.0	1825	63.5638	0.825	0.3075	1.2465	0.825	0.8021	0.1607	0.0466
61.52	74.0	1850	63.5654	0.82	0.3058	1.2425	0.82	0.7963	0.1552	0.0468
61.52	75.0	1875	63.5654	0.825	0.3041	1.2439	0.825	0.8021	0.1563	0.0466
61.52	76.0	1900	63.5499	0.83	0.3018	1.2432	0.83	0.8082	0.1541	0.0463
61.52	77.0	1925	63.5563	0.825	0.3059	1.2385	0.825	0.8021	0.1570	0.0466
61.52	78.0	1950	63.5524	0.825	0.3045	1.2364	0.825	0.8021	0.1524	0.0464
61.52	79.0	1975	63.5507	0.825	0.3064	1.2344	0.825	0.8021	0.1523	0.0463
61.4257	80.0	2000	63.5531	0.825	0.3062	1.2266	0.825	0.8035	0.1625	0.0463
61.4257	81.0	2025	63.5486	0.825	0.3029	1.1850	0.825	0.8024	0.1506	0.0463
61.4257	82.0	2050	63.5479	0.82	0.3081	1.2269	0.82	0.7963	0.1588	0.0458
61.4257	83.0	2075	63.5444	0.835	0.3029	1.1721	0.835	0.8139	0.1475	0.0461
61.4257	84.0	2100	63.5435	0.835	0.3047	1.2306	0.835	0.8171	0.1529	0.0464
61.4257	85.0	2125	63.5393	0.83	0.3058	1.2255	0.83	0.8081	0.1462	0.0464
61.4257	86.0	2150	63.5437	0.835	0.3048	1.2254	0.835	0.8171	0.1481	0.0464
61.4257	87.0	2175	63.5463	0.83	0.3039	1.1549	0.83	0.8115	0.1562	0.0463
61.4257	88.0	2200	63.5408	0.835	0.3055	1.2211	0.835	0.8187	0.1485	0.0462
61.4257	89.0	2225	63.5477	0.825	0.3054	1.1541	0.825	0.8024	0.1521	0.0463
61.4257	90.0	2250	63.5383	0.83	0.3051	1.1577	0.83	0.8095	0.1532	0.0463
61.4257	91.0	2275	63.5466	0.84	0.3057	1.1583	0.8400	0.8244	0.1516	0.0458
61.4257	92.0	2300	63.5447	0.835	0.3049	1.1518	0.835	0.8188	0.1615	0.0462
61.4257	93.0	2325	63.5327	0.84	0.3044	1.1540	0.8400	0.8244	0.1508	0.0459
61.4257	94.0	2350	63.5392	0.84	0.3046	1.1506	0.8400	0.8244	0.1569	0.0459
61.4257	95.0	2375	63.5305	0.835	0.3050	1.1520	0.835	0.8188	0.1571	0.0457
61.4257	96.0	2400	63.5413	0.835	0.3042	1.1494	0.835	0.8188	0.1571	0.0461
61.4257	97.0	2425	63.5387	0.835	0.3047	1.1489	0.835	0.8188	0.1652	0.0461
61.4257	98.0	2450	63.5383	0.84	0.3046	1.1503	0.8400	0.8244	0.1568	0.0458
61.4257	99.0	2475	63.5374	0.835	0.3045	1.1489	0.835	0.8188	0.1570	0.0456
61.3919	100.0	2500	63.5396	0.84	0.3043	1.1495	0.8400	0.8244	0.1568	0.0457

Framework versions

Transformers 4.26.1
Pytorch 1.13.1.post200
Datasets 2.9.0
Tokenizers 0.13.2

jordyvl/60-tiny_tobacco3482_hint_

60-tiny_tobacco3482_hint_

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions