81-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: 63.0326
Accuracy: 0.85
Brier Loss: 0.2647
Nll: 1.1178
F1 Micro: 0.85
F1 Macro: 0.8409
Ece: 0.1296
Aurc: 0.0380

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	66.3150	0.26	0.8706	4.9002	0.26	0.1920	0.2904	0.7808
No log	2.0	50	65.5312	0.54	0.5948	2.8544	0.54	0.4535	0.2938	0.2564
No log	3.0	75	65.0613	0.675	0.4403	1.6101	0.675	0.6134	0.2276	0.1328
No log	4.0	100	64.7775	0.765	0.3800	1.5512	0.765	0.7486	0.2394	0.1111
No log	5.0	125	64.4695	0.8	0.3065	1.7477	0.8000	0.7658	0.1814	0.0713
No log	6.0	150	64.3179	0.78	0.3286	1.7786	0.78	0.7409	0.1555	0.0636
No log	7.0	175	64.0707	0.77	0.3516	1.6165	0.7700	0.7179	0.1677	0.0927
No log	8.0	200	63.8878	0.775	0.3677	1.5868	0.775	0.7675	0.1922	0.0867
No log	9.0	225	64.2141	0.715	0.4341	2.1772	0.715	0.7074	0.2273	0.1130
No log	10.0	250	63.8037	0.775	0.3649	1.4471	0.775	0.7221	0.1744	0.0766
No log	11.0	275	64.2412	0.655	0.5396	2.0448	0.655	0.6532	0.2692	0.1670
No log	12.0	300	63.8514	0.73	0.4372	1.9358	0.7300	0.6997	0.2194	0.1008
No log	13.0	325	63.4352	0.815	0.2888	1.6462	0.815	0.7858	0.1490	0.0469
No log	14.0	350	63.5101	0.795	0.3262	1.3794	0.795	0.7816	0.1599	0.0648
No log	15.0	375	63.6365	0.785	0.3704	1.4766	0.785	0.7627	0.1691	0.0781
No log	16.0	400	63.7245	0.73	0.4145	1.7447	0.7300	0.7261	0.2087	0.0917
No log	17.0	425	63.4312	0.795	0.3148	1.4363	0.795	0.7868	0.1665	0.0636
No log	18.0	450	63.7070	0.835	0.2915	1.4455	0.835	0.8078	0.1542	0.0570
No log	19.0	475	63.3600	0.81	0.2997	1.3326	0.81	0.7878	0.1602	0.0534
62.5692	20.0	500	63.4339	0.81	0.3158	1.2051	0.81	0.7809	0.1541	0.0506
62.5692	21.0	525	63.3477	0.805	0.3088	1.2575	0.805	0.7943	0.1570	0.0489
62.5692	22.0	550	63.3362	0.83	0.2911	1.3409	0.83	0.8104	0.1440	0.0514
62.5692	23.0	575	63.3897	0.805	0.3004	1.2505	0.805	0.7833	0.1595	0.0453
62.5692	24.0	600	63.3475	0.8	0.3185	1.1190	0.8000	0.7750	0.1626	0.0486
62.5692	25.0	625	63.4552	0.805	0.3470	1.2483	0.805	0.7904	0.1818	0.0652
62.5692	26.0	650	63.4364	0.79	0.3453	1.1298	0.79	0.7798	0.1827	0.0651
62.5692	27.0	675	63.3001	0.83	0.2899	1.4329	0.83	0.8141	0.1370	0.0466
62.5692	28.0	700	63.1848	0.85	0.2514	1.2175	0.85	0.8354	0.1220	0.0396
62.5692	29.0	725	63.2303	0.835	0.2744	1.4886	0.835	0.8143	0.1389	0.0532
62.5692	30.0	750	63.2275	0.84	0.2774	1.0426	0.8400	0.8405	0.1325	0.0632
62.5692	31.0	775	63.1341	0.835	0.2597	1.1066	0.835	0.8083	0.1170	0.0510
62.5692	32.0	800	63.2045	0.81	0.2970	1.0432	0.81	0.8028	0.1561	0.0552
62.5692	33.0	825	63.1898	0.82	0.2952	1.0189	0.82	0.8060	0.1440	0.0521
62.5692	34.0	850	63.1330	0.835	0.2760	1.0095	0.835	0.8265	0.1356	0.0542
62.5692	35.0	875	63.1572	0.84	0.2834	1.0174	0.8400	0.8234	0.1337	0.0508
62.5692	36.0	900	63.1922	0.835	0.2894	1.0102	0.835	0.8233	0.1469	0.0500
62.5692	37.0	925	63.1305	0.83	0.2818	1.0146	0.83	0.8172	0.1387	0.0510
62.5692	38.0	950	63.1902	0.815	0.2865	1.0101	0.815	0.8016	0.1500	0.0516
62.5692	39.0	975	63.1835	0.825	0.2851	1.0177	0.825	0.8162	0.1436	0.0496
61.2333	40.0	1000	63.1741	0.84	0.2783	1.0160	0.8400	0.8266	0.1275	0.0510
61.2333	41.0	1025	63.1755	0.835	0.2756	1.0117	0.835	0.8192	0.1447	0.0483
61.2333	42.0	1050	63.1281	0.83	0.2820	1.0142	0.83	0.8169	0.1415	0.0466
61.2333	43.0	1075	63.1697	0.85	0.2675	0.9929	0.85	0.8358	0.1423	0.0484
61.2333	44.0	1100	63.1141	0.835	0.2767	1.0005	0.835	0.8237	0.1293	0.0481
61.2333	45.0	1125	63.1441	0.85	0.2638	1.0023	0.85	0.8383	0.1335	0.0471
61.2333	46.0	1150	63.1221	0.84	0.2745	0.9981	0.8400	0.8308	0.1271	0.0451
61.2333	47.0	1175	63.1140	0.845	0.2654	0.9891	0.845	0.8317	0.1351	0.0458
61.2333	48.0	1200	63.1056	0.845	0.2654	1.0016	0.845	0.8351	0.1364	0.0458
61.2333	49.0	1225	63.0906	0.83	0.2713	1.0042	0.83	0.8221	0.1455	0.0449
61.2333	50.0	1250	63.0942	0.835	0.2633	1.0003	0.835	0.8314	0.1397	0.0452
61.2333	51.0	1275	63.0929	0.84	0.2641	0.9957	0.8400	0.8359	0.1340	0.0440
61.2333	52.0	1300	63.0913	0.83	0.2646	1.0040	0.83	0.8242	0.1422	0.0440
61.2333	53.0	1325	63.1152	0.83	0.2754	0.9985	0.83	0.8250	0.1416	0.0447
61.2333	54.0	1350	63.0923	0.835	0.2649	0.9997	0.835	0.8278	0.1356	0.0426
61.2333	55.0	1375	63.0720	0.83	0.2686	0.9988	0.83	0.8243	0.1396	0.0431
61.2333	56.0	1400	63.0627	0.83	0.2636	1.0713	0.83	0.8243	0.1369	0.0427
61.2333	57.0	1425	63.0742	0.835	0.2692	1.0572	0.835	0.8305	0.1391	0.0425
61.2333	58.0	1450	63.0910	0.84	0.2639	1.0727	0.8400	0.8334	0.1320	0.0432
61.2333	59.0	1475	63.1015	0.84	0.2648	1.1382	0.8400	0.8354	0.1331	0.0423
61.0482	60.0	1500	63.0557	0.835	0.2655	1.0688	0.835	0.8293	0.1333	0.0420
61.0482	61.0	1525	63.0590	0.835	0.2655	1.1378	0.835	0.8315	0.1425	0.0416
61.0482	62.0	1550	63.0732	0.845	0.2661	1.0565	0.845	0.8381	0.1404	0.0413
61.0482	63.0	1575	63.0972	0.855	0.2659	1.1274	0.855	0.8501	0.1424	0.0416
61.0482	64.0	1600	63.0528	0.84	0.2694	1.1315	0.8400	0.8330	0.1355	0.0418
61.0482	65.0	1625	63.0625	0.835	0.2683	1.1336	0.835	0.8281	0.1373	0.0411
61.0482	66.0	1650	63.0512	0.835	0.2747	1.1250	0.835	0.8242	0.1436	0.0410
61.0482	67.0	1675	63.0634	0.85	0.2671	1.1270	0.85	0.8397	0.1376	0.0419
61.0482	68.0	1700	63.0609	0.835	0.2717	1.1311	0.835	0.8295	0.1365	0.0411
61.0482	69.0	1725	63.0513	0.835	0.2707	1.1261	0.835	0.8223	0.1461	0.0412
61.0482	70.0	1750	63.0510	0.845	0.2712	1.1219	0.845	0.8396	0.1369	0.0411
61.0482	71.0	1775	63.0530	0.845	0.2688	1.1244	0.845	0.8364	0.1403	0.0412
61.0482	72.0	1800	63.0456	0.84	0.2665	1.1204	0.8400	0.8293	0.1341	0.0400
61.0482	73.0	1825	63.0459	0.845	0.2680	1.1244	0.845	0.8360	0.1430	0.0398
61.0482	74.0	1850	63.0773	0.85	0.2684	1.1291	0.85	0.8440	0.1386	0.0410
61.0482	75.0	1875	63.0497	0.85	0.2664	1.1285	0.85	0.8388	0.1293	0.0392
61.0482	76.0	1900	63.0483	0.845	0.2695	1.1256	0.845	0.8352	0.1440	0.0409
61.0482	77.0	1925	63.0352	0.845	0.2680	1.1229	0.845	0.8352	0.1420	0.0398
61.0482	78.0	1950	63.0291	0.845	0.2701	1.1225	0.845	0.8352	0.1342	0.0394
61.0482	79.0	1975	63.0508	0.85	0.2695	1.1224	0.85	0.8388	0.1418	0.0399
60.9704	80.0	2000	63.0510	0.85	0.2708	1.1169	0.85	0.8388	0.1376	0.0394
60.9704	81.0	2025	63.0460	0.85	0.2648	1.1205	0.85	0.8440	0.1357	0.0397
60.9704	82.0	2050	63.0505	0.845	0.2697	1.1148	0.845	0.8352	0.1464	0.0392
60.9704	83.0	2075	63.0425	0.845	0.2651	1.1229	0.845	0.8352	0.1396	0.0389
60.9704	84.0	2100	63.0398	0.845	0.2664	1.1197	0.845	0.8337	0.1330	0.0388
60.9704	85.0	2125	63.0355	0.845	0.2667	1.1192	0.845	0.8360	0.1307	0.0387
60.9704	86.0	2150	63.0386	0.85	0.2649	1.1223	0.85	0.8409	0.1279	0.0379
60.9704	87.0	2175	63.0405	0.85	0.2642	1.1218	0.85	0.8409	0.1437	0.0378
60.9704	88.0	2200	63.0363	0.85	0.2667	1.1165	0.85	0.8388	0.1320	0.0390
60.9704	89.0	2225	63.0456	0.845	0.2644	1.1180	0.845	0.8352	0.1354	0.0381
60.9704	90.0	2250	63.0343	0.845	0.2656	1.1159	0.845	0.8337	0.1390	0.0385
60.9704	91.0	2275	63.0391	0.85	0.2654	1.1194	0.85	0.8409	0.1389	0.0380
60.9704	92.0	2300	63.0354	0.85	0.2665	1.1203	0.85	0.8409	0.1419	0.0377
60.9704	93.0	2325	63.0272	0.845	0.2650	1.1166	0.845	0.8381	0.1370	0.0381
60.9704	94.0	2350	63.0313	0.85	0.2647	1.1181	0.85	0.8409	0.1322	0.0380
60.9704	95.0	2375	63.0220	0.845	0.2658	1.1197	0.845	0.8357	0.1311	0.0378
60.9704	96.0	2400	63.0345	0.85	0.2639	1.1179	0.85	0.8409	0.1284	0.0381
60.9704	97.0	2425	63.0330	0.845	0.2651	1.1163	0.845	0.8352	0.1391	0.0382
60.9704	98.0	2450	63.0302	0.85	0.2646	1.1182	0.85	0.8409	0.1311	0.0380
60.9704	99.0	2475	63.0287	0.85	0.2646	1.1175	0.85	0.8409	0.1250	0.0380
60.9392	100.0	2500	63.0326	0.85	0.2647	1.1178	0.85	0.8409	0.1296	0.0380

Framework versions

Transformers 4.26.1
Pytorch 1.13.1.post200
Datasets 2.9.0
Tokenizers 0.13.2

jordyvl/81-tiny_tobacco3482_hint_

81-tiny_tobacco3482_hint_

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions