81-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: 230.0462
Accuracy: 0.845
Brier Loss: 0.2451
Nll: 1.1250
F1 Micro: 0.845
F1 Macro: 0.8350
Ece: 0.1115
Aurc: 0.0383

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	238.2967	0.275	0.8789	6.2995	0.275	0.1849	0.2898	0.4790
No log	2.0	100	236.7866	0.41	0.7214	3.0130	0.41	0.2599	0.2544	0.3985
No log	3.0	150	235.3687	0.59	0.5690	2.4483	0.59	0.4507	0.2464	0.2336
No log	4.0	200	234.7639	0.65	0.4658	2.3318	0.65	0.4996	0.2365	0.1474
No log	5.0	250	234.0798	0.675	0.4692	2.3351	0.675	0.5158	0.2359	0.1335
No log	6.0	300	234.2002	0.68	0.4267	2.1570	0.68	0.5724	0.1990	0.1124
No log	7.0	350	234.1154	0.665	0.4874	2.9269	0.665	0.5631	0.2249	0.1259
No log	8.0	400	233.2390	0.77	0.3224	1.8874	0.7700	0.6850	0.1754	0.0695
No log	9.0	450	233.1396	0.755	0.3583	2.1139	0.755	0.7060	0.1573	0.0802
234.3515	10.0	500	233.5895	0.705	0.4653	3.2740	0.705	0.6813	0.2183	0.1180
234.3515	11.0	550	233.0481	0.745	0.3803	2.7330	0.745	0.6697	0.1730	0.0936
234.3515	12.0	600	232.9339	0.78	0.3319	1.8094	0.78	0.7416	0.1700	0.0660
234.3515	13.0	650	233.0181	0.755	0.3182	1.9630	0.755	0.7596	0.1772	0.0823
234.3515	14.0	700	232.4934	0.815	0.2659	1.3869	0.815	0.7749	0.1512	0.0511
234.3515	15.0	750	232.6922	0.79	0.3361	1.8793	0.79	0.7671	0.1633	0.0926
234.3515	16.0	800	232.3391	0.815	0.2785	1.5488	0.815	0.8038	0.1476	0.0549
234.3515	17.0	850	232.5412	0.775	0.3100	2.0192	0.775	0.7675	0.1617	0.0502
234.3515	18.0	900	232.5307	0.785	0.3211	1.6793	0.785	0.7541	0.1679	0.0676
234.3515	19.0	950	232.3055	0.825	0.2862	1.5993	0.825	0.8204	0.1443	0.0620
231.365	20.0	1000	232.5636	0.78	0.3431	2.6048	0.78	0.7835	0.1658	0.0847
231.365	21.0	1050	231.8748	0.845	0.2737	1.4332	0.845	0.8454	0.1616	0.0602
231.365	22.0	1100	232.1032	0.825	0.2557	1.6240	0.825	0.8290	0.1422	0.0442
231.365	23.0	1150	231.7278	0.835	0.2740	1.6260	0.835	0.8246	0.1541	0.0613
231.365	24.0	1200	231.9350	0.84	0.2636	1.6028	0.8400	0.8410	0.1472	0.0471
231.365	25.0	1250	231.6054	0.82	0.2753	1.3835	0.82	0.8090	0.1526	0.0478
231.365	26.0	1300	231.6516	0.84	0.2737	1.3849	0.8400	0.8354	0.1492	0.0543
231.365	27.0	1350	231.5453	0.84	0.2536	1.1384	0.8400	0.8265	0.1363	0.0431
231.365	28.0	1400	231.4833	0.85	0.2292	1.1206	0.85	0.8350	0.1250	0.0362
231.365	29.0	1450	231.3722	0.815	0.2856	1.1706	0.815	0.8038	0.1606	0.0486
230.3328	30.0	1500	231.3517	0.84	0.2608	1.3387	0.8400	0.8382	0.1366	0.0483
230.3328	31.0	1550	231.3705	0.815	0.2724	1.2558	0.815	0.7992	0.1448	0.0463
230.3328	32.0	1600	231.4319	0.84	0.2588	1.0691	0.8400	0.8301	0.1317	0.0435
230.3328	33.0	1650	231.2119	0.86	0.2323	1.1693	0.8600	0.8609	0.1229	0.0470
230.3328	34.0	1700	231.2836	0.83	0.2477	1.1294	0.83	0.8201	0.1398	0.0406
230.3328	35.0	1750	231.2669	0.845	0.2569	1.1508	0.845	0.8369	0.1357	0.0449
230.3328	36.0	1800	231.0634	0.85	0.2422	1.0830	0.85	0.8404	0.1372	0.0419
230.3328	37.0	1850	231.1141	0.85	0.2398	1.0879	0.85	0.8455	0.1347	0.0437
230.3328	38.0	1900	231.0520	0.815	0.2626	1.2325	0.815	0.8077	0.1304	0.0478
230.3328	39.0	1950	230.9089	0.83	0.2507	1.1245	0.83	0.8258	0.1347	0.0487
229.6409	40.0	2000	231.0532	0.86	0.2207	1.1258	0.8600	0.8464	0.1206	0.0428
229.6409	41.0	2050	230.9307	0.855	0.2350	1.1326	0.855	0.8498	0.1284	0.0401
229.6409	42.0	2100	230.8493	0.86	0.2306	1.1075	0.8600	0.8548	0.1441	0.0417
229.6409	43.0	2150	230.7516	0.87	0.2198	1.0312	0.87	0.8587	0.1184	0.0404
229.6409	44.0	2200	230.8540	0.85	0.2485	1.1724	0.85	0.8444	0.1425	0.0451
229.6409	45.0	2250	230.7995	0.86	0.2284	1.1183	0.8600	0.8490	0.1284	0.0368
229.6409	46.0	2300	230.7162	0.825	0.2701	1.1206	0.825	0.8159	0.1390	0.0437
229.6409	47.0	2350	230.5593	0.855	0.2341	1.2242	0.855	0.8459	0.1226	0.0392
229.6409	48.0	2400	230.6472	0.86	0.2377	1.0233	0.8600	0.8558	0.1319	0.0354
229.6409	49.0	2450	230.7080	0.84	0.2548	1.1208	0.8400	0.8294	0.1484	0.0426
229.142	50.0	2500	230.5862	0.845	0.2543	1.2129	0.845	0.8322	0.1358	0.0415
229.142	51.0	2550	230.6550	0.845	0.2462	1.0937	0.845	0.8333	0.1306	0.0411
229.142	52.0	2600	230.5789	0.835	0.2595	1.1393	0.835	0.8249	0.1369	0.0428
229.142	53.0	2650	230.5895	0.85	0.2519	1.0185	0.85	0.8447	0.1263	0.0439
229.142	54.0	2700	230.4955	0.86	0.2402	1.0837	0.8600	0.8590	0.1382	0.0394
229.142	55.0	2750	230.4579	0.84	0.2560	1.1514	0.8400	0.8312	0.1439	0.0431
229.142	56.0	2800	230.5190	0.845	0.2527	1.2868	0.845	0.8334	0.1342	0.0406
229.142	57.0	2850	230.4989	0.84	0.2536	1.0785	0.8400	0.8278	0.1377	0.0397
229.142	58.0	2900	230.4445	0.84	0.2416	1.2104	0.8400	0.8340	0.1339	0.0383
229.142	59.0	2950	230.3159	0.87	0.2211	1.1138	0.87	0.8630	0.1171	0.0394
228.7442	60.0	3000	230.3405	0.835	0.2603	1.1327	0.835	0.8273	0.1422	0.0416
228.7442	61.0	3050	230.3892	0.845	0.2395	1.1619	0.845	0.8308	0.1263	0.0377
228.7442	62.0	3100	230.3728	0.84	0.2478	1.0053	0.8400	0.8291	0.1208	0.0389
228.7442	63.0	3150	230.2486	0.845	0.2545	1.1989	0.845	0.8438	0.1364	0.0434
228.7442	64.0	3200	230.2353	0.84	0.2512	1.1260	0.8400	0.8330	0.1191	0.0430
228.7442	65.0	3250	230.2041	0.845	0.2429	1.1529	0.845	0.8384	0.1343	0.0401
228.7442	66.0	3300	230.2439	0.845	0.2511	1.1125	0.845	0.8366	0.1367	0.0430
228.7442	67.0	3350	230.1510	0.85	0.2471	1.0528	0.85	0.8457	0.1414	0.0402
228.7442	68.0	3400	230.2274	0.85	0.2455	1.1150	0.85	0.8397	0.1311	0.0427
228.7442	69.0	3450	230.2165	0.845	0.2524	1.0517	0.845	0.8421	0.1312	0.0410
228.4757	70.0	3500	230.1976	0.835	0.2600	1.0845	0.835	0.8258	0.1353	0.0410
228.4757	71.0	3550	230.1062	0.85	0.2487	1.1447	0.85	0.8410	0.1297	0.0427
228.4757	72.0	3600	229.9867	0.835	0.2584	1.0641	0.835	0.8273	0.1236	0.0440
228.4757	73.0	3650	230.1918	0.845	0.2411	1.1521	0.845	0.8363	0.1373	0.0389
228.4757	74.0	3700	230.0781	0.85	0.2524	1.0980	0.85	0.8390	0.1298	0.0409
228.4757	75.0	3750	230.1432	0.835	0.2554	1.0967	0.835	0.8230	0.1227	0.0407
228.4757	76.0	3800	230.1512	0.84	0.2535	1.0945	0.8400	0.8295	0.1321	0.0422
228.4757	77.0	3850	230.0682	0.84	0.2502	1.0301	0.8400	0.8370	0.1312	0.0403
228.4757	78.0	3900	230.0357	0.835	0.2521	1.1572	0.835	0.8293	0.1244	0.0412
228.4757	79.0	3950	230.1252	0.845	0.2509	1.0961	0.845	0.8381	0.1273	0.0409
228.2815	80.0	4000	230.0584	0.845	0.2539	1.0795	0.845	0.8363	0.1235	0.0432
228.2815	81.0	4050	229.9967	0.85	0.2427	1.1156	0.85	0.8382	0.1184	0.0394
228.2815	82.0	4100	230.0755	0.84	0.2563	1.0833	0.8400	0.8302	0.1295	0.0406
228.2815	83.0	4150	230.0798	0.845	0.2477	1.1713	0.845	0.8385	0.1259	0.0427
228.2815	84.0	4200	230.0299	0.84	0.2477	1.0907	0.8400	0.8260	0.1213	0.0383
228.2815	85.0	4250	230.0568	0.845	0.2483	1.0763	0.845	0.8350	0.1238	0.0409
228.2815	86.0	4300	230.0743	0.85	0.2464	1.0549	0.85	0.8418	0.1271	0.0398
228.2815	87.0	4350	230.0061	0.845	0.2505	1.1585	0.845	0.8350	0.1312	0.0375
228.2815	88.0	4400	229.9674	0.845	0.2478	1.0763	0.845	0.8346	0.1394	0.0410
228.2815	89.0	4450	229.9697	0.85	0.2451	1.0833	0.85	0.8406	0.1324	0.0364
228.1298	90.0	4500	230.0305	0.845	0.2496	1.1008	0.845	0.8350	0.1308	0.0395
228.1298	91.0	4550	229.9740	0.845	0.2495	1.0605	0.845	0.8350	0.1309	0.0413
228.1298	92.0	4600	229.9962	0.85	0.2497	1.1193	0.85	0.8408	0.1294	0.0399
228.1298	93.0	4650	229.9740	0.85	0.2491	1.0496	0.85	0.8383	0.1270	0.0390
228.1298	94.0	4700	229.9698	0.84	0.2516	1.0644	0.8400	0.8295	0.1430	0.0398
228.1298	95.0	4750	229.9247	0.845	0.2516	1.1705	0.845	0.8350	0.1227	0.0371
228.1298	96.0	4800	230.0451	0.84	0.2507	1.0970	0.8400	0.8285	0.1368	0.0391
228.1298	97.0	4850	229.9402	0.85	0.2495	1.1427	0.85	0.8430	0.1379	0.0413
228.1298	98.0	4900	230.0130	0.84	0.2532	1.0964	0.8400	0.8285	0.1412	0.0389
228.1298	99.0	4950	230.0200	0.845	0.2482	1.0916	0.845	0.8317	0.1190	0.0392
228.0583	100.0	5000	230.0462	0.845	0.2451	1.1250	0.845	0.8350	0.1115	0.0383

Framework versions

Transformers 4.26.1
Pytorch 1.13.1.post200
Datasets 2.9.0
Tokenizers 0.13.2

jordyvl/81-tiny_tobacco3482_og_simkd

81-tiny_tobacco3482_og_simkd

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions