300-tiny_tobacco3482_og_simkd

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: 1030.4594
Accuracy: 0.845
Brier Loss: 0.2507
Nll: 1.6602
F1 Micro: 0.845
F1 Macro: 0.8273
Ece: 0.1440
Aurc: 0.0504

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: 16
eval_batch_size: 16
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	50	1063.5994	0.215	0.8548	4.5394	0.2150	0.0866	0.2467	0.5619
No log	2.0	100	1058.8120	0.53	0.6920	2.3234	0.53	0.4420	0.3178	0.2841
No log	3.0	150	1054.1038	0.63	0.5596	2.2480	0.63	0.5709	0.3050	0.1733
No log	4.0	200	1049.1915	0.67	0.4903	2.6916	0.67	0.5866	0.2481	0.1460
No log	5.0	250	1049.7939	0.71	0.4663	1.9364	0.7100	0.6359	0.2438	0.1365
No log	6.0	300	1048.1417	0.69	0.4830	2.1830	0.69	0.6309	0.2518	0.1379
No log	7.0	350	1048.1290	0.725	0.4265	2.1076	0.7250	0.6383	0.2114	0.1172
No log	8.0	400	1044.3457	0.775	0.3807	1.9343	0.775	0.7467	0.2407	0.0982
No log	9.0	450	1046.8309	0.67	0.4528	2.7967	0.67	0.6493	0.2279	0.1225
1185.4053	10.0	500	1044.5997	0.76	0.3748	2.0448	0.76	0.7066	0.1889	0.0965
1185.4053	11.0	550	1046.6857	0.745	0.3799	2.1911	0.745	0.7195	0.2129	0.0936
1185.4053	12.0	600	1036.3792	0.69	0.4258	2.9605	0.69	0.6744	0.1946	0.1255
1185.4053	13.0	650	1042.4764	0.715	0.4055	2.8193	0.715	0.7326	0.1879	0.0969
1185.4053	14.0	700	1040.1765	0.765	0.3811	2.5360	0.765	0.7497	0.2079	0.0957
1185.4053	15.0	750	1045.1747	0.655	0.4732	3.4740	0.655	0.6577	0.2284	0.1180
1185.4053	16.0	800	1035.7695	0.785	0.3445	2.7152	0.785	0.7564	0.1876	0.1176
1185.4053	17.0	850	1037.5389	0.755	0.3937	2.8771	0.755	0.7418	0.2153	0.1097
1185.4053	18.0	900	1045.9521	0.775	0.3902	2.4149	0.775	0.7261	0.2202	0.1157
1185.4053	19.0	950	1039.2958	0.81	0.3229	1.8751	0.81	0.8025	0.1808	0.0855
1173.9102	20.0	1000	1037.2432	0.765	0.3557	2.4940	0.765	0.7378	0.1849	0.0929
1173.9102	21.0	1050	1042.5205	0.785	0.3416	2.1312	0.785	0.7641	0.2139	0.0791
1173.9102	22.0	1100	1031.1548	0.785	0.3667	2.6195	0.785	0.7555	0.1945	0.1464
1173.9102	23.0	1150	1038.0391	0.815	0.3085	2.4661	0.815	0.7769	0.1814	0.0923
1173.9102	24.0	1200	1033.7769	0.75	0.3758	2.5229	0.75	0.7341	0.1910	0.1272
1173.9102	25.0	1250	1040.0878	0.82	0.2923	2.2061	0.82	0.7882	0.1810	0.0600
1173.9102	26.0	1300	1041.4139	0.84	0.2750	2.0378	0.8400	0.8224	0.1849	0.0645
1173.9102	27.0	1350	1033.1660	0.835	0.2818	1.9645	0.835	0.8018	0.1471	0.0771
1173.9102	28.0	1400	1036.1976	0.82	0.2973	2.2955	0.82	0.7993	0.1691	0.0755
1173.9102	29.0	1450	1035.8601	0.845	0.2740	2.1097	0.845	0.8171	0.1696	0.0637
1169.7248	30.0	1500	1033.2889	0.82	0.3046	2.2615	0.82	0.7985	0.1820	0.0910
1169.7248	31.0	1550	1030.7528	0.85	0.2549	1.8264	0.85	0.8265	0.1459	0.0798
1169.7248	32.0	1600	1030.7676	0.835	0.2708	1.9641	0.835	0.8127	0.1646	0.0648
1169.7248	33.0	1650	1033.2319	0.84	0.2689	1.9015	0.8400	0.8091	0.1652	0.0741
1169.7248	34.0	1700	1039.8257	0.82	0.2974	2.2209	0.82	0.7998	0.1767	0.0687
1169.7248	35.0	1750	1035.8896	0.865	0.2331	1.8113	0.865	0.8458	0.1535	0.0674
1169.7248	36.0	1800	1038.4092	0.82	0.3014	2.1408	0.82	0.8013	0.1708	0.0844
1169.7248	37.0	1850	1034.9086	0.84	0.2722	1.8873	0.8400	0.8233	0.1509	0.0689
1169.7248	38.0	1900	1034.8866	0.84	0.2751	2.1269	0.8400	0.8196	0.1647	0.0781
1169.7248	39.0	1950	1031.3431	0.815	0.2934	2.0669	0.815	0.7844	0.1669	0.0839
1167.1226	40.0	2000	1031.2334	0.84	0.2766	2.1016	0.8400	0.8322	0.1580	0.0847
1167.1226	41.0	2050	1033.4176	0.84	0.2725	1.7005	0.8400	0.8114	0.1473	0.0731
1167.1226	42.0	2100	1032.7366	0.83	0.2871	2.1337	0.83	0.8037	0.1592	0.0852
1167.1226	43.0	2150	1032.4486	0.845	0.2683	2.0076	0.845	0.8218	0.1645	0.0852
1167.1226	44.0	2200	1035.2195	0.84	0.2660	1.8355	0.8400	0.8147	0.1670	0.0643
1167.1226	45.0	2250	1034.2225	0.84	0.2556	1.8249	0.8400	0.8022	0.1622	0.0598
1167.1226	46.0	2300	1032.7642	0.845	0.2534	1.6936	0.845	0.8166	0.1438	0.0639
1167.1226	47.0	2350	1036.8080	0.83	0.2765	1.7153	0.83	0.8076	0.1638	0.0600
1167.1226	48.0	2400	1033.1261	0.865	0.2290	1.7616	0.865	0.8577	0.1402	0.0536
1167.1226	49.0	2450	1036.7234	0.84	0.2656	1.7275	0.8400	0.8093	0.1459	0.0597
1165.4459	50.0	2500	1026.7501	0.85	0.2643	2.0250	0.85	0.8276	0.1577	0.0791
1165.4459	51.0	2550	1033.3628	0.825	0.2804	1.7547	0.825	0.7943	0.1592	0.0589
1165.4459	52.0	2600	1033.1439	0.845	0.2433	1.8201	0.845	0.8156	0.1370	0.0559
1165.4459	53.0	2650	1031.2720	0.85	0.2602	1.8013	0.85	0.8276	0.1550	0.0637
1165.4459	54.0	2700	1031.4177	0.845	0.2625	1.7994	0.845	0.8295	0.1514	0.0590
1165.4459	55.0	2750	1029.6421	0.84	0.2534	1.5053	0.8400	0.8163	0.1512	0.0541
1165.4459	56.0	2800	1032.8274	0.85	0.2475	1.7805	0.85	0.8405	0.1384	0.0554
1165.4459	57.0	2850	1030.2781	0.835	0.2608	1.6087	0.835	0.8059	0.1570	0.0585
1165.4459	58.0	2900	1032.2584	0.85	0.2434	1.6753	0.85	0.8311	0.1389	0.0606
1165.4459	59.0	2950	1031.4507	0.845	0.2590	1.7351	0.845	0.8291	0.1464	0.0556
1163.88	60.0	3000	1031.8136	0.85	0.2461	1.8215	0.85	0.8329	0.1438	0.0578
1163.88	61.0	3050	1032.6718	0.86	0.2410	1.6261	0.8600	0.8465	0.1483	0.0566
1163.88	62.0	3100	1028.8600	0.855	0.2488	1.6111	0.855	0.8358	0.1421	0.0578
1163.88	63.0	3150	1029.1522	0.845	0.2787	1.8558	0.845	0.8135	0.1539	0.0591
1163.88	64.0	3200	1026.7474	0.85	0.2457	1.9070	0.85	0.8280	0.1411	0.0570
1163.88	65.0	3250	1032.3363	0.845	0.2595	1.7872	0.845	0.8169	0.1435	0.0616
1163.88	66.0	3300	1031.6006	0.855	0.2474	1.7063	0.855	0.8357	0.1357	0.0521
1163.88	67.0	3350	1031.1060	0.855	0.2420	1.5538	0.855	0.8401	0.1357	0.0562
1163.88	68.0	3400	1032.2750	0.855	0.2443	1.7408	0.855	0.8455	0.1348	0.0520
1163.88	69.0	3450	1031.3723	0.84	0.2523	1.6058	0.8400	0.8211	0.1356	0.0550
1162.7836	70.0	3500	1032.2831	0.855	0.2539	1.6918	0.855	0.8386	0.1406	0.0605
1162.7836	71.0	3550	1028.9554	0.85	0.2479	1.6822	0.85	0.8290	0.1395	0.0567
1162.7836	72.0	3600	1031.4127	0.855	0.2495	1.7386	0.855	0.8386	0.1365	0.0564
1162.7836	73.0	3650	1031.2170	0.86	0.2450	1.7208	0.8600	0.8454	0.1337	0.0551
1162.7836	74.0	3700	1030.2428	0.86	0.2458	1.7499	0.8600	0.8454	0.1490	0.0525
1162.7836	75.0	3750	1031.7998	0.86	0.2442	1.6081	0.8600	0.8384	0.1373	0.0569
1162.7836	76.0	3800	1030.5450	0.85	0.2469	1.5548	0.85	0.8354	0.1380	0.0563
1162.7836	77.0	3850	1031.4061	0.855	0.2383	1.6571	0.855	0.8376	0.1313	0.0549
1162.7836	78.0	3900	1033.5166	0.85	0.2554	1.5379	0.85	0.8279	0.1400	0.0572
1162.7836	79.0	3950	1033.4709	0.855	0.2496	1.5471	0.855	0.8437	0.1287	0.0524
1162.105	80.0	4000	1031.8766	0.85	0.2453	1.6612	0.85	0.8325	0.1388	0.0556
1162.105	81.0	4050	1030.5789	0.855	0.2441	1.7341	0.855	0.8356	0.1259	0.0535
1162.105	82.0	4100	1033.6682	0.85	0.2517	1.6263	0.85	0.8340	0.1443	0.0513
1162.105	83.0	4150	1031.2977	0.855	0.2425	1.6676	0.855	0.8427	0.1334	0.0537
1162.105	84.0	4200	1027.3252	0.86	0.2456	1.7430	0.8600	0.8515	0.1326	0.0542
1162.105	85.0	4250	1030.4331	0.855	0.2454	1.7415	0.855	0.8439	0.1507	0.0501
1162.105	86.0	4300	1031.9957	0.845	0.2553	1.6819	0.845	0.8322	0.1463	0.0497
1162.105	87.0	4350	1031.3456	0.845	0.2543	1.7632	0.845	0.8240	0.1503	0.0556
1162.105	88.0	4400	1030.5261	0.845	0.2522	1.7459	0.845	0.8243	0.1450	0.0482
1162.105	89.0	4450	1028.4374	0.845	0.2507	1.6968	0.845	0.8243	0.1356	0.0538
1161.3639	90.0	4500	1030.4808	0.845	0.2501	1.6386	0.845	0.8240	0.1442	0.0531
1161.3639	91.0	4550	1032.4968	0.845	0.2583	1.7297	0.845	0.8235	0.1329	0.0519
1161.3639	92.0	4600	1031.8092	0.85	0.2523	1.6594	0.85	0.8340	0.1411	0.0519
1161.3639	93.0	4650	1031.8936	0.85	0.2539	1.7723	0.85	0.8366	0.1392	0.0475
1161.3639	94.0	4700	1031.4855	0.855	0.2525	1.6232	0.855	0.8448	0.1477	0.0508
1161.3639	95.0	4750	1027.4985	0.855	0.2502	1.8411	0.855	0.8448	0.1442	0.0510
1161.3639	96.0	4800	1031.0199	0.85	0.2500	1.7087	0.85	0.8353	0.1342	0.0488
1161.3639	97.0	4850	1031.1750	0.84	0.2592	1.7172	0.8400	0.8152	0.1407	0.0530
1161.3639	98.0	4900	1032.8660	0.845	0.2541	1.6367	0.845	0.8273	0.1352	0.0488
1161.3639	99.0	4950	1029.1611	0.855	0.2509	1.6982	0.855	0.8448	0.1485	0.0525
1161.2033	100.0	5000	1030.4594	0.845	0.2507	1.6602	0.845	0.8273	0.1440	0.0504

Framework versions

Transformers 4.26.1
Pytorch 1.13.1.post200
Datasets 2.9.0
Tokenizers 0.13.2

jordyvl/300-tiny_tobacco3482_og_simkd

300-tiny_tobacco3482_og_simkd

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions