39-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: 65.8239
Accuracy: 0.84
Brier Loss: 0.2807
Nll: 1.1327
F1 Micro: 0.8400
F1 Macro: 0.8280
Ece: 0.1437
Aurc: 0.0472

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	69.1264	0.26	0.8707	4.9002	0.26	0.1920	0.3064	0.7815
No log	2.0	50	68.3319	0.545	0.5960	2.8558	0.545	0.4562	0.2850	0.2564
No log	3.0	75	67.8627	0.68	0.4406	1.6064	0.68	0.6157	0.2543	0.1333
No log	4.0	100	67.5797	0.75	0.3829	1.5484	0.75	0.7343	0.2220	0.1152
No log	5.0	125	67.2608	0.8	0.3072	1.7491	0.8000	0.7573	0.1809	0.0698
No log	6.0	150	67.0950	0.78	0.3169	1.7708	0.78	0.7441	0.1576	0.0607
No log	7.0	175	66.9178	0.755	0.3812	1.6929	0.755	0.6848	0.1899	0.1043
No log	8.0	200	66.7335	0.75	0.3763	1.7649	0.75	0.7399	0.1870	0.0806
No log	9.0	225	66.4371	0.805	0.3036	1.4686	0.805	0.7896	0.1378	0.0545
No log	10.0	250	66.6823	0.75	0.3924	1.8808	0.75	0.6665	0.1964	0.0829
No log	11.0	275	66.6079	0.775	0.3570	1.8872	0.775	0.7442	0.1811	0.0839
No log	12.0	300	66.4364	0.765	0.3689	1.6981	0.765	0.7550	0.1909	0.0732
No log	13.0	325	66.1317	0.785	0.3346	1.4062	0.785	0.7823	0.1753	0.0572
No log	14.0	350	66.5182	0.73	0.4453	1.4431	0.7300	0.7208	0.2310	0.0985
No log	15.0	375	66.5154	0.775	0.3769	1.4897	0.775	0.7398	0.1875	0.0821
No log	16.0	400	66.4191	0.8	0.3315	1.5327	0.8000	0.7771	0.1651	0.0685
No log	17.0	425	66.2374	0.765	0.3520	1.5388	0.765	0.7401	0.1767	0.0760
No log	18.0	450	66.2010	0.805	0.3320	1.4280	0.805	0.7721	0.1756	0.0684
No log	19.0	475	66.0335	0.85	0.2625	1.3549	0.85	0.8352	0.1431	0.0430
65.4034	20.0	500	66.2213	0.815	0.3213	1.3912	0.815	0.7955	0.1645	0.0579
65.4034	21.0	525	66.2647	0.77	0.3656	1.3241	0.7700	0.7743	0.1899	0.0755
65.4034	22.0	550	66.1220	0.86	0.2684	1.2459	0.8600	0.8354	0.1327	0.0473
65.4034	23.0	575	66.1615	0.85	0.2623	1.3400	0.85	0.8231	0.1291	0.0448
65.4034	24.0	600	66.2114	0.825	0.3130	1.4118	0.825	0.8122	0.1565	0.0498
65.4034	25.0	625	66.1048	0.835	0.2704	1.3571	0.835	0.8196	0.1405	0.0450
65.4034	26.0	650	65.9832	0.825	0.2990	1.1514	0.825	0.8253	0.1603	0.0423
65.4034	27.0	675	66.2567	0.805	0.3307	1.3509	0.805	0.8022	0.1699	0.0634
65.4034	28.0	700	66.0668	0.82	0.3172	1.1445	0.82	0.7973	0.1538	0.0419
65.4034	29.0	725	66.2254	0.81	0.3252	1.3290	0.81	0.8011	0.1659	0.0523
65.4034	30.0	750	65.9643	0.84	0.2697	1.2052	0.8400	0.8245	0.1319	0.0425
65.4034	31.0	775	66.3419	0.81	0.3249	1.2772	0.81	0.7969	0.1700	0.0612
65.4034	32.0	800	66.0324	0.825	0.3003	1.3138	0.825	0.8000	0.1584	0.0445
65.4034	33.0	825	66.3326	0.82	0.3336	1.2983	0.82	0.7826	0.1754	0.0590
65.4034	34.0	850	66.1374	0.825	0.3061	1.5645	0.825	0.8012	0.1500	0.0459
65.4034	35.0	875	66.2310	0.815	0.3207	1.5607	0.815	0.7939	0.1712	0.0646
65.4034	36.0	900	66.0388	0.84	0.2873	1.1966	0.8400	0.8327	0.1456	0.0585
65.4034	37.0	925	66.0520	0.835	0.2958	1.2728	0.835	0.8180	0.1508	0.0477
65.4034	38.0	950	65.9916	0.84	0.2783	1.1635	0.8400	0.8233	0.1398	0.0438
65.4034	39.0	975	65.9391	0.845	0.2743	1.2660	0.845	0.8289	0.1396	0.0458
64.0802	40.0	1000	65.9291	0.845	0.2762	1.3335	0.845	0.8259	0.1373	0.0430
64.0802	41.0	1025	65.8559	0.85	0.2686	1.3432	0.85	0.8338	0.1345	0.0428
64.0802	42.0	1050	65.8612	0.845	0.2772	1.2679	0.845	0.8255	0.1389	0.0431
64.0802	43.0	1075	65.8953	0.84	0.2742	1.2614	0.8400	0.8227	0.1408	0.0435
64.0802	44.0	1100	65.8569	0.835	0.2769	1.2730	0.835	0.8199	0.1426	0.0432
64.0802	45.0	1125	65.8610	0.84	0.2769	1.2622	0.8400	0.8248	0.1485	0.0425
64.0802	46.0	1150	65.8237	0.845	0.2729	1.1920	0.845	0.8334	0.1462	0.0432
64.0802	47.0	1175	65.8416	0.845	0.2785	1.1826	0.845	0.8317	0.1376	0.0431
64.0802	48.0	1200	65.8452	0.845	0.2817	1.1876	0.845	0.8317	0.1417	0.0441
64.0802	49.0	1225	65.8394	0.845	0.2750	1.1993	0.845	0.8309	0.1315	0.0419
64.0802	50.0	1250	65.8527	0.84	0.2796	1.1860	0.8400	0.8279	0.1410	0.0432
64.0802	51.0	1275	65.8286	0.845	0.2749	1.1977	0.845	0.8333	0.1444	0.0428
64.0802	52.0	1300	65.8296	0.83	0.2779	1.1926	0.83	0.8171	0.1382	0.0435
64.0802	53.0	1325	65.8121	0.83	0.2779	1.1955	0.83	0.8155	0.1387	0.0436
64.0802	54.0	1350	65.8361	0.825	0.2769	1.1909	0.825	0.8095	0.1435	0.0419
64.0802	55.0	1375	65.8370	0.83	0.2816	1.1925	0.83	0.8171	0.1416	0.0435
64.0802	56.0	1400	65.8301	0.825	0.2763	1.1908	0.825	0.8101	0.1393	0.0439
64.0802	57.0	1425	65.8301	0.82	0.2791	1.1881	0.82	0.8040	0.1443	0.0440
64.0802	58.0	1450	65.8324	0.83	0.2754	1.1938	0.83	0.8198	0.1387	0.0460
64.0802	59.0	1475	65.8407	0.825	0.2818	1.1893	0.825	0.8138	0.1393	0.0439
63.8765	60.0	1500	65.8236	0.84	0.2782	1.1871	0.8400	0.8290	0.1512	0.0449
63.8765	61.0	1525	65.8198	0.825	0.2846	1.1752	0.825	0.8138	0.1505	0.0438
63.8765	62.0	1550	65.8243	0.83	0.2796	1.1753	0.83	0.8196	0.1480	0.0445
63.8765	63.0	1575	65.8495	0.835	0.2781	1.1766	0.835	0.8257	0.1353	0.0451
63.8765	64.0	1600	65.8204	0.835	0.2833	1.1752	0.835	0.8239	0.1400	0.0447
63.8765	65.0	1625	65.8374	0.835	0.2800	1.1829	0.835	0.8239	0.1474	0.0441
63.8765	66.0	1650	65.8433	0.83	0.2855	1.1678	0.83	0.8148	0.1498	0.0444
63.8765	67.0	1675	65.8259	0.835	0.2820	1.1725	0.835	0.8257	0.1518	0.0457
63.8765	68.0	1700	65.8443	0.83	0.2841	1.1652	0.83	0.8196	0.1491	0.0457
63.8765	69.0	1725	65.8255	0.835	0.2849	1.1620	0.835	0.8247	0.1499	0.0460
63.8765	70.0	1750	65.8421	0.83	0.2870	1.1681	0.83	0.8196	0.1418	0.0475
63.8765	71.0	1775	65.8402	0.835	0.2839	1.1614	0.835	0.8230	0.1359	0.0466
63.8765	72.0	1800	65.8224	0.84	0.2831	1.1555	0.8400	0.8280	0.1467	0.0459
63.8765	73.0	1825	65.8233	0.84	0.2824	1.1578	0.8400	0.8280	0.1428	0.0465
63.8765	74.0	1850	65.8299	0.84	0.2814	1.1574	0.8400	0.8280	0.1469	0.0465
63.8765	75.0	1875	65.8309	0.835	0.2790	1.1575	0.835	0.8219	0.1407	0.0465
63.8765	76.0	1900	65.8199	0.84	0.2789	1.1496	0.8400	0.8280	0.1437	0.0460
63.8765	77.0	1925	65.8222	0.84	0.2828	1.1520	0.8400	0.8280	0.1539	0.0461
63.8765	78.0	1950	65.8312	0.84	0.2801	1.1459	0.8400	0.8280	0.1354	0.0458
63.8765	79.0	1975	65.8253	0.84	0.2836	1.1448	0.8400	0.8280	0.1542	0.0465
63.7964	80.0	2000	65.8332	0.84	0.2839	1.1408	0.8400	0.8280	0.1486	0.0462
63.7964	81.0	2025	65.8316	0.84	0.2818	1.1419	0.8400	0.8280	0.1430	0.0460
63.7964	82.0	2050	65.8238	0.84	0.2824	1.1387	0.8400	0.8280	0.1411	0.0452
63.7964	83.0	2075	65.8294	0.84	0.2786	1.1410	0.8400	0.8280	0.1539	0.0469
63.7964	84.0	2100	65.8267	0.84	0.2818	1.1391	0.8400	0.8280	0.1463	0.0471
63.7964	85.0	2125	65.8222	0.84	0.2814	1.1401	0.8400	0.8280	0.1463	0.0470
63.7964	86.0	2150	65.8264	0.84	0.2776	1.1380	0.8400	0.8280	0.1359	0.0460
63.7964	87.0	2175	65.8228	0.84	0.2781	1.1366	0.8400	0.8280	0.1468	0.0460
63.7964	88.0	2200	65.8229	0.84	0.2832	1.1367	0.8400	0.8280	0.1455	0.0476
63.7964	89.0	2225	65.8271	0.84	0.2792	1.1376	0.8400	0.8280	0.1598	0.0467
63.7964	90.0	2250	65.8234	0.84	0.2830	1.1352	0.8400	0.8280	0.1427	0.0474
63.7964	91.0	2275	65.8309	0.84	0.2804	1.1352	0.8400	0.8280	0.1426	0.0467
63.7964	92.0	2300	65.8305	0.84	0.2796	1.1345	0.8400	0.8280	0.1438	0.0466
63.7964	93.0	2325	65.8155	0.84	0.2808	1.1347	0.8400	0.8280	0.1499	0.0471
63.7964	94.0	2350	65.8218	0.84	0.2803	1.1336	0.8400	0.8280	0.1487	0.0473
63.7964	95.0	2375	65.8152	0.84	0.2812	1.1334	0.8400	0.8280	0.1441	0.0466
63.7964	96.0	2400	65.8230	0.84	0.2801	1.1344	0.8400	0.8280	0.1488	0.0472
63.7964	97.0	2425	65.8206	0.84	0.2808	1.1328	0.8400	0.8280	0.1490	0.0472
63.7964	98.0	2450	65.8221	0.84	0.2807	1.1332	0.8400	0.8280	0.1438	0.0474
63.7964	99.0	2475	65.8207	0.84	0.2809	1.1326	0.8400	0.8280	0.1446	0.0472
63.7613	100.0	2500	65.8239	0.84	0.2807	1.1327	0.8400	0.8280	0.1437	0.0472

Framework versions

Transformers 4.26.1
Pytorch 1.13.1.post200
Datasets 2.9.0
Tokenizers 0.13.2

jordyvl/39-tiny_tobacco3482_hint_

39-tiny_tobacco3482_hint_

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions