vit-small_tobacco3482_og_simkd_

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: 212.9178
Accuracy: 0.855
Brier Loss: 0.2563
Nll: 1.4722
F1 Micro: 0.855
F1 Macro: 0.8333
Ece: 0.1253
Aurc: 0.0422

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	219.2297	0.145	0.8895	6.5562	0.145	0.0567	0.2109	0.7661
No log	2.0	50	217.9786	0.49	0.6839	2.2035	0.49	0.4113	0.3237	0.3097
No log	3.0	75	216.6085	0.565	0.5671	1.7658	0.565	0.4471	0.2471	0.2239
No log	4.0	100	216.0210	0.68	0.4722	1.8682	0.68	0.5586	0.2322	0.1557
No log	5.0	125	215.5695	0.68	0.4668	1.9385	0.68	0.5570	0.2289	0.1440
No log	6.0	150	215.3762	0.745	0.3963	2.1043	0.745	0.6608	0.1949	0.0976
No log	7.0	175	214.8964	0.745	0.3675	1.7226	0.745	0.6693	0.1765	0.0949
No log	8.0	200	215.0440	0.735	0.3838	1.9180	0.735	0.6935	0.2056	0.0958
No log	9.0	225	214.7017	0.775	0.3466	1.5816	0.775	0.6897	0.1756	0.0661
No log	10.0	250	214.6309	0.775	0.3505	1.6245	0.775	0.7604	0.1828	0.0763
No log	11.0	275	214.6275	0.735	0.4314	2.3367	0.735	0.7342	0.2234	0.1203
No log	12.0	300	214.5664	0.75	0.3769	1.7889	0.75	0.7420	0.1873	0.1070
No log	13.0	325	214.6764	0.735	0.4425	2.3533	0.735	0.7404	0.2267	0.1508
No log	14.0	350	214.5261	0.805	0.3093	1.8504	0.805	0.7870	0.1732	0.0580
No log	15.0	375	214.4932	0.79	0.3255	1.4649	0.79	0.7575	0.1796	0.0543
No log	16.0	400	214.3134	0.85	0.2467	1.4769	0.85	0.8388	0.1149	0.0513
No log	17.0	425	214.3825	0.82	0.2845	1.4858	0.82	0.8014	0.1445	0.0540
No log	18.0	450	214.2077	0.85	0.2681	1.4891	0.85	0.8406	0.1462	0.0684
No log	19.0	475	214.1675	0.845	0.2623	1.5311	0.845	0.8329	0.1414	0.0485
220.0633	20.0	500	214.1433	0.84	0.2663	1.5269	0.8400	0.8182	0.1302	0.0562
220.0633	21.0	525	214.0829	0.805	0.3293	2.0021	0.805	0.8019	0.1710	0.0833
220.0633	22.0	550	213.9282	0.84	0.2586	1.5127	0.8400	0.8205	0.1397	0.0453
220.0633	23.0	575	213.9303	0.87	0.2260	1.4450	0.87	0.8552	0.1205	0.0365
220.0633	24.0	600	213.9140	0.84	0.2620	1.5244	0.8400	0.8161	0.1462	0.0490
220.0633	25.0	625	213.7616	0.86	0.2306	1.5288	0.8600	0.8409	0.1215	0.0361
220.0633	26.0	650	213.7738	0.845	0.2431	1.5303	0.845	0.8271	0.1335	0.0443
220.0633	27.0	675	213.8470	0.85	0.2427	1.3459	0.85	0.8275	0.1296	0.0445
220.0633	28.0	700	213.7198	0.85	0.2381	1.3868	0.85	0.8328	0.1267	0.0424
220.0633	29.0	725	213.6302	0.855	0.2293	1.4191	0.855	0.8361	0.1157	0.0394
220.0633	30.0	750	213.6385	0.85	0.2424	1.5410	0.85	0.8334	0.1339	0.0464
220.0633	31.0	775	213.6397	0.865	0.2234	1.4012	0.865	0.8464	0.1226	0.0402
220.0633	32.0	800	213.6658	0.86	0.2271	1.3863	0.8600	0.8470	0.1164	0.0354
220.0633	33.0	825	213.6526	0.85	0.2448	1.5357	0.85	0.8292	0.1214	0.0397
220.0633	34.0	850	213.5407	0.855	0.2282	1.3470	0.855	0.8405	0.1245	0.0393
220.0633	35.0	875	213.6166	0.83	0.2624	1.4288	0.83	0.8102	0.1415	0.0458
220.0633	36.0	900	213.5887	0.84	0.2613	1.3928	0.8400	0.8135	0.1298	0.0442
220.0633	37.0	925	213.4976	0.845	0.2338	1.3784	0.845	0.8244	0.1319	0.0355
220.0633	38.0	950	213.4554	0.85	0.2374	1.3680	0.85	0.8323	0.1192	0.0385
220.0633	39.0	975	213.4758	0.845	0.2319	1.4895	0.845	0.8274	0.1185	0.0385
217.7609	40.0	1000	213.4440	0.845	0.2432	1.3737	0.845	0.8265	0.1310	0.0415
217.7609	41.0	1025	213.4492	0.845	0.2385	1.4970	0.845	0.8297	0.1207	0.0373
217.7609	42.0	1050	213.4319	0.85	0.2384	1.3580	0.85	0.8276	0.1250	0.0383
217.7609	43.0	1075	213.3094	0.855	0.2287	1.4375	0.855	0.8334	0.1188	0.0353
217.7609	44.0	1100	213.3809	0.845	0.2514	1.4969	0.845	0.8250	0.1318	0.0435
217.7609	45.0	1125	213.3981	0.85	0.2478	1.6052	0.85	0.8287	0.1268	0.0408
217.7609	46.0	1150	213.3004	0.86	0.2292	1.3632	0.8600	0.8430	0.1180	0.0355
217.7609	47.0	1175	213.3041	0.86	0.2363	1.3407	0.8600	0.8444	0.1235	0.0359
217.7609	48.0	1200	213.2955	0.845	0.2462	1.5071	0.845	0.8179	0.1253	0.0396
217.7609	49.0	1225	213.2531	0.85	0.2433	1.2946	0.85	0.8277	0.1270	0.0392
217.7609	50.0	1250	213.2612	0.845	0.2378	1.2852	0.845	0.8193	0.1281	0.0361
217.7609	51.0	1275	213.2246	0.855	0.2370	1.5829	0.855	0.8393	0.1234	0.0357
217.7609	52.0	1300	213.1795	0.845	0.2431	1.4923	0.845	0.8300	0.1280	0.0372
217.7609	53.0	1325	213.2721	0.855	0.2467	1.5096	0.855	0.8333	0.1248	0.0385
217.7609	54.0	1350	213.1976	0.85	0.2453	1.4167	0.85	0.8275	0.1240	0.0384
217.7609	55.0	1375	213.2822	0.845	0.2430	1.4438	0.845	0.8193	0.1283	0.0396
217.7609	56.0	1400	213.1443	0.85	0.2479	1.5246	0.85	0.8277	0.1304	0.0389
217.7609	57.0	1425	213.1679	0.85	0.2455	1.4468	0.85	0.8291	0.1224	0.0387
217.7609	58.0	1450	213.1116	0.85	0.2467	1.4372	0.85	0.8287	0.1269	0.0378
217.7609	59.0	1475	213.1005	0.85	0.2490	1.4214	0.85	0.8271	0.1316	0.0392
217.1217	60.0	1500	213.1516	0.855	0.2425	1.4600	0.855	0.8343	0.1316	0.0369
217.1217	61.0	1525	213.1205	0.855	0.2458	1.4436	0.855	0.8303	0.1197	0.0409
217.1217	62.0	1550	213.1318	0.85	0.2488	1.4405	0.85	0.8275	0.1304	0.0378
217.1217	63.0	1575	213.0243	0.855	0.2521	1.5810	0.855	0.8328	0.1341	0.0447
217.1217	64.0	1600	213.1191	0.84	0.2567	1.4478	0.8400	0.8185	0.1292	0.0436
217.1217	65.0	1625	213.0329	0.855	0.2528	1.3910	0.855	0.8333	0.1311	0.0404
217.1217	66.0	1650	212.9868	0.85	0.2525	1.4652	0.85	0.8275	0.1226	0.0408
217.1217	67.0	1675	213.0856	0.84	0.2561	1.4601	0.8400	0.8178	0.1367	0.0419
217.1217	68.0	1700	213.0379	0.845	0.2544	1.5222	0.845	0.8216	0.1362	0.0426
217.1217	69.0	1725	213.0535	0.835	0.2606	1.5085	0.835	0.8093	0.1346	0.0445
217.1217	70.0	1750	213.0247	0.85	0.2530	1.4349	0.85	0.8274	0.1373	0.0427
217.1217	71.0	1775	213.0161	0.855	0.2510	1.4529	0.855	0.8333	0.1212	0.0411
217.1217	72.0	1800	213.0249	0.845	0.2494	1.4511	0.845	0.8229	0.1358	0.0412
217.1217	73.0	1825	213.0014	0.85	0.2548	1.4435	0.85	0.8264	0.1277	0.0390
217.1217	74.0	1850	213.0011	0.845	0.2527	1.3719	0.845	0.8206	0.1360	0.0379
217.1217	75.0	1875	213.0240	0.845	0.2576	1.4072	0.845	0.8221	0.1284	0.0425
217.1217	76.0	1900	212.9793	0.845	0.2534	1.4026	0.845	0.8212	0.1241	0.0404
217.1217	77.0	1925	212.9800	0.85	0.2514	1.5023	0.85	0.8271	0.1279	0.0407
217.1217	78.0	1950	212.9125	0.845	0.2564	1.4258	0.845	0.8211	0.1298	0.0427
217.1217	79.0	1975	212.9454	0.85	0.2527	1.5227	0.85	0.8271	0.1279	0.0423
216.765	80.0	2000	212.9475	0.845	0.2551	1.5025	0.845	0.8206	0.1311	0.0423
216.765	81.0	2025	212.9739	0.84	0.2567	1.5305	0.8400	0.8162	0.1294	0.0431
216.765	82.0	2050	212.9351	0.855	0.2526	1.5373	0.855	0.8339	0.1277	0.0401
216.765	83.0	2075	213.0053	0.845	0.2560	1.4724	0.845	0.8228	0.1341	0.0417
216.765	84.0	2100	212.9326	0.845	0.2568	1.5217	0.845	0.8206	0.1303	0.0472
216.765	85.0	2125	212.9555	0.855	0.2537	1.5265	0.855	0.8339	0.1233	0.0416
216.765	86.0	2150	212.9121	0.85	0.2534	1.5224	0.85	0.8280	0.1283	0.0398
216.765	87.0	2175	212.8850	0.845	0.2551	1.4480	0.845	0.8221	0.1328	0.0412
216.765	88.0	2200	212.9121	0.855	0.2518	1.5069	0.855	0.8339	0.1234	0.0404
216.765	89.0	2225	212.9327	0.845	0.2517	1.4532	0.845	0.8206	0.1231	0.0401
216.765	90.0	2250	212.9305	0.85	0.2542	1.4506	0.85	0.8271	0.1374	0.0398
216.765	91.0	2275	212.9274	0.85	0.2567	1.5045	0.85	0.8280	0.1297	0.0419
216.765	92.0	2300	212.8962	0.85	0.2545	1.4956	0.85	0.8280	0.1261	0.0405
216.765	93.0	2325	212.9133	0.845	0.2567	1.5274	0.845	0.8212	0.1291	0.0431
216.765	94.0	2350	212.8708	0.85	0.2576	1.4410	0.85	0.8280	0.1302	0.0410
216.765	95.0	2375	212.9661	0.855	0.2546	1.3988	0.855	0.8339	0.1248	0.0404
216.765	96.0	2400	212.9099	0.855	0.2547	1.5096	0.855	0.8333	0.1256	0.0402
216.765	97.0	2425	212.9668	0.85	0.2549	1.5337	0.85	0.8271	0.1289	0.0390
216.765	98.0	2450	212.9587	0.845	0.2545	1.5161	0.845	0.8215	0.1304	0.0412
216.765	99.0	2475	212.9395	0.855	0.2554	1.4606	0.855	0.8333	0.1253	0.0410
216.6085	100.0	2500	212.9178	0.855	0.2563	1.4722	0.855	0.8333	0.1253	0.0422

Framework versions

Transformers 4.26.1
Pytorch 1.13.1.post200
Datasets 2.9.0
Tokenizers 0.13.2

jordyvl/vit-small_tobacco3482_og_simkd_

vit-small_tobacco3482_og_simkd_

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions