dit-base_tobacco-small_tobacco3482_kd_MSE

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: 0.7746
Accuracy: 0.81
Brier Loss: 0.2775
Nll: 1.1981
F1 Micro: 0.81
F1 Macro: 0.7980
Ece: 0.1403
Aurc: 0.0500

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	7.0083	0.14	0.9250	5.4655	0.14	0.1468	0.2824	0.8920
No log	2.0	14	5.8247	0.35	0.7844	3.6804	0.35	0.2240	0.2541	0.5601
No log	3.0	21	4.1788	0.49	0.6140	1.8305	0.49	0.4563	0.2512	0.2825
No log	4.0	28	2.7911	0.66	0.4534	1.6541	0.66	0.5604	0.2299	0.1475
No log	5.0	35	2.3354	0.74	0.3892	1.8678	0.74	0.6851	0.2104	0.0989
No log	6.0	42	1.9675	0.73	0.3585	1.3943	0.7300	0.6822	0.1846	0.0930
No log	7.0	49	1.7187	0.79	0.3190	1.3921	0.79	0.7510	0.1739	0.0760
No log	8.0	56	1.6507	0.77	0.3469	1.3682	0.7700	0.7289	0.1834	0.0851
No log	9.0	63	1.2713	0.79	0.3040	1.4042	0.79	0.7622	0.1505	0.0540
No log	10.0	70	1.1461	0.805	0.2852	1.3953	0.805	0.7849	0.1371	0.0522
No log	11.0	77	1.1328	0.81	0.2713	1.3113	0.81	0.7901	0.1371	0.0442
No log	12.0	84	1.2818	0.8	0.3192	1.2680	0.8000	0.7808	0.1674	0.0725
No log	13.0	91	1.0493	0.805	0.2767	1.2512	0.805	0.7846	0.1451	0.0535
No log	14.0	98	0.9657	0.815	0.2802	1.1796	0.815	0.7965	0.1680	0.0487
No log	15.0	105	0.9910	0.82	0.2695	1.3658	0.82	0.8000	0.1400	0.0475
No log	16.0	112	0.9828	0.81	0.2823	1.3175	0.81	0.7974	0.1390	0.0549
No log	17.0	119	0.9279	0.8	0.2815	1.3727	0.8000	0.7882	0.1599	0.0454
No log	18.0	126	1.0076	0.805	0.2929	1.2999	0.805	0.7825	0.1480	0.0562
No log	19.0	133	0.9524	0.82	0.2705	1.3029	0.82	0.8122	0.1481	0.0454
No log	20.0	140	1.0584	0.795	0.3010	1.3019	0.795	0.7699	0.1669	0.0650
No log	21.0	147	0.9390	0.805	0.2775	1.4073	0.805	0.7888	0.1211	0.0513
No log	22.0	154	0.9857	0.81	0.2895	1.2894	0.81	0.7879	0.1469	0.0548
No log	23.0	161	0.9137	0.795	0.2809	1.4461	0.795	0.7872	0.1528	0.0472
No log	24.0	168	0.8545	0.815	0.2844	1.2582	0.815	0.7981	0.1466	0.0484
No log	25.0	175	0.8860	0.81	0.2766	1.4525	0.81	0.8010	0.1241	0.0457
No log	26.0	182	0.8624	0.83	0.2813	1.1993	0.83	0.8222	0.1536	0.0512
No log	27.0	189	0.9119	0.805	0.2894	1.4164	0.805	0.7869	0.1576	0.0519
No log	28.0	196	0.9072	0.82	0.2753	1.2927	0.82	0.8149	0.1292	0.0514
No log	29.0	203	0.8428	0.8	0.2805	1.3065	0.8000	0.7820	0.1368	0.0502
No log	30.0	210	0.8696	0.81	0.2858	1.2825	0.81	0.7989	0.1454	0.0524
No log	31.0	217	0.8542	0.8	0.2861	1.2029	0.8000	0.7766	0.1412	0.0496
No log	32.0	224	0.8576	0.805	0.2896	1.3371	0.805	0.7814	0.1513	0.0515
No log	33.0	231	0.8615	0.8	0.2859	1.2347	0.8000	0.7826	0.1473	0.0522
No log	34.0	238	0.8474	0.805	0.2807	1.3510	0.805	0.7946	0.1493	0.0524
No log	35.0	245	0.9058	0.79	0.3035	1.2005	0.79	0.7768	0.1497	0.0553
No log	36.0	252	0.8461	0.805	0.2897	1.2770	0.805	0.7906	0.1599	0.0513
No log	37.0	259	0.8461	0.805	0.2962	1.1989	0.805	0.7912	0.1527	0.0533
No log	38.0	266	0.8646	0.815	0.2817	1.3653	0.815	0.8031	0.1355	0.0499
No log	39.0	273	0.8306	0.8	0.2905	1.1852	0.8000	0.7862	0.1528	0.0549
No log	40.0	280	0.8561	0.815	0.2838	1.2577	0.815	0.8005	0.1431	0.0544
No log	41.0	287	0.8236	0.805	0.2836	1.2093	0.805	0.7925	0.1376	0.0490
No log	42.0	294	0.8221	0.805	0.2853	1.1929	0.805	0.7805	0.1397	0.0524
No log	43.0	301	0.7834	0.815	0.2666	1.2720	0.815	0.8006	0.1316	0.0496
No log	44.0	308	0.8022	0.8	0.2839	1.2009	0.8000	0.7870	0.1457	0.0514
No log	45.0	315	0.8009	0.81	0.2735	1.3505	0.81	0.7970	0.1359	0.0494
No log	46.0	322	0.8029	0.81	0.2775	1.1956	0.81	0.7983	0.1476	0.0509
No log	47.0	329	0.7979	0.82	0.2818	1.2005	0.82	0.8049	0.1466	0.0488
No log	48.0	336	0.7763	0.815	0.2784	1.1905	0.815	0.7970	0.1358	0.0512
No log	49.0	343	0.7917	0.81	0.2802	1.2136	0.81	0.7989	0.1429	0.0486
No log	50.0	350	0.8223	0.825	0.2809	1.1860	0.825	0.8042	0.1567	0.0520
No log	51.0	357	0.7952	0.82	0.2747	1.2074	0.82	0.8078	0.1377	0.0484
No log	52.0	364	0.7868	0.825	0.2714	1.2850	0.825	0.8170	0.1371	0.0476
No log	53.0	371	0.8111	0.805	0.2869	1.1892	0.805	0.7954	0.1467	0.0524
No log	54.0	378	0.7739	0.81	0.2755	1.1946	0.81	0.7953	0.1567	0.0486
No log	55.0	385	0.7930	0.825	0.2825	1.2000	0.825	0.8087	0.1546	0.0518
No log	56.0	392	0.7826	0.815	0.2789	1.1953	0.815	0.8031	0.1353	0.0514
No log	57.0	399	0.7716	0.82	0.2714	1.3115	0.82	0.8079	0.1207	0.0470
No log	58.0	406	0.8036	0.815	0.2878	1.1875	0.815	0.7945	0.1469	0.0531
No log	59.0	413	0.7714	0.82	0.2722	1.2787	0.82	0.8128	0.1264	0.0467
No log	60.0	420	0.7671	0.825	0.2720	1.2722	0.825	0.8136	0.1378	0.0476
No log	61.0	427	0.7885	0.815	0.2834	1.1798	0.815	0.8007	0.1480	0.0526
No log	62.0	434	0.7621	0.82	0.2706	1.3459	0.82	0.8102	0.1156	0.0482
No log	63.0	441	0.7691	0.81	0.2797	1.1379	0.81	0.7959	0.1429	0.0506
No log	64.0	448	0.7699	0.81	0.2776	1.1964	0.81	0.7974	0.1473	0.0494
No log	65.0	455	0.7693	0.82	0.2739	1.2089	0.82	0.8106	0.1390	0.0481
No log	66.0	462	0.7891	0.81	0.2805	1.1989	0.81	0.7927	0.1530	0.0513
No log	67.0	469	0.7806	0.82	0.2798	1.2033	0.82	0.8068	0.1408	0.0485
No log	68.0	476	0.7877	0.82	0.2815	1.1896	0.82	0.8054	0.1376	0.0501
No log	69.0	483	0.7649	0.825	0.2731	1.1567	0.825	0.8155	0.1371	0.0479
No log	70.0	490	0.7740	0.82	0.2764	1.1929	0.82	0.8107	0.1250	0.0511
No log	71.0	497	0.7657	0.82	0.2744	1.2762	0.82	0.8068	0.1374	0.0488
0.4804	72.0	504	0.7887	0.805	0.2839	1.1851	0.805	0.7914	0.1524	0.0513
0.4804	73.0	511	0.7662	0.815	0.2759	1.1973	0.815	0.8010	0.1395	0.0496
0.4804	74.0	518	0.7706	0.825	0.2742	1.2020	0.825	0.8196	0.1398	0.0492
0.4804	75.0	525	0.7780	0.815	0.2802	1.1881	0.815	0.7970	0.1392	0.0505
0.4804	76.0	532	0.7731	0.825	0.2745	1.2695	0.825	0.8152	0.1548	0.0485
0.4804	77.0	539	0.7743	0.825	0.2762	1.2039	0.825	0.8109	0.1326	0.0490
0.4804	78.0	546	0.7782	0.805	0.2792	1.2001	0.805	0.7905	0.1381	0.0506
0.4804	79.0	553	0.7786	0.81	0.2807	1.1929	0.81	0.7980	0.1394	0.0505
0.4804	80.0	560	0.7759	0.82	0.2772	1.1973	0.82	0.8081	0.1296	0.0494
0.4804	81.0	567	0.7703	0.82	0.2758	1.2069	0.82	0.8096	0.1405	0.0491
0.4804	82.0	574	0.7749	0.81	0.2777	1.1996	0.81	0.7980	0.1502	0.0501
0.4804	83.0	581	0.7768	0.815	0.2777	1.2009	0.815	0.8052	0.1237	0.0496
0.4804	84.0	588	0.7761	0.815	0.2778	1.1986	0.815	0.8008	0.1333	0.0495
0.4804	85.0	595	0.7771	0.815	0.2780	1.1984	0.815	0.8008	0.1335	0.0497
0.4804	86.0	602	0.7755	0.81	0.2777	1.1987	0.81	0.7980	0.1327	0.0501
0.4804	87.0	609	0.7749	0.81	0.2776	1.1974	0.81	0.7980	0.1261	0.0499
0.4804	88.0	616	0.7746	0.815	0.2776	1.1981	0.815	0.8052	0.1238	0.0497
0.4804	89.0	623	0.7744	0.81	0.2776	1.1981	0.81	0.7980	0.1283	0.0500
0.4804	90.0	630	0.7743	0.81	0.2774	1.1987	0.81	0.7980	0.1346	0.0499
0.4804	91.0	637	0.7741	0.81	0.2774	1.1981	0.81	0.7980	0.1379	0.0499
0.4804	92.0	644	0.7742	0.81	0.2774	1.1982	0.81	0.7980	0.1403	0.0499
0.4804	93.0	651	0.7745	0.81	0.2775	1.1982	0.81	0.7980	0.1403	0.0500
0.4804	94.0	658	0.7746	0.81	0.2776	1.1978	0.81	0.7980	0.1316	0.0500
0.4804	95.0	665	0.7745	0.81	0.2775	1.1982	0.81	0.7980	0.1380	0.0499
0.4804	96.0	672	0.7746	0.81	0.2775	1.1981	0.81	0.7980	0.1315	0.0500
0.4804	97.0	679	0.7746	0.81	0.2775	1.1981	0.81	0.7980	0.1403	0.0500
0.4804	98.0	686	0.7746	0.81	0.2775	1.1981	0.81	0.7980	0.1403	0.0500
0.4804	99.0	693	0.7746	0.81	0.2775	1.1981	0.81	0.7980	0.1403	0.0500
0.4804	100.0	700	0.7746	0.81	0.2775	1.1981	0.81	0.7980	0.1403	0.0500

Framework versions

Transformers 4.36.0.dev0
Pytorch 2.2.0.dev20231112+cu118
Datasets 2.14.5
Tokenizers 0.14.1

jordyvl/dit-base_tobacco-small_tobacco3482_kd_MSE

dit-base_tobacco-small_tobacco3482_kd_MSE

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions