39-tiny_tobacco3482

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: 0.0652
Accuracy: 0.68
Brier Loss: 0.8297
Nll: 1.9403
F1 Micro: 0.68
F1 Macro: 0.5899
Ece: 0.5728
Aurc: 0.1321

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	0.0877	0.11	0.8983	8.4701	0.11	0.0377	0.1765	0.9065
No log	2.0	50	0.0811	0.09	0.8984	6.6861	0.09	0.0364	0.1641	0.9033
No log	3.0	75	0.0803	0.165	0.8981	7.2209	0.165	0.0520	0.2115	0.7480
No log	4.0	100	0.0799	0.17	0.8977	7.1709	0.17	0.0713	0.2184	0.6439
No log	5.0	125	0.0794	0.15	0.8970	7.1655	0.15	0.0553	0.2075	0.6278
No log	6.0	150	0.0786	0.225	0.8959	6.1378	0.225	0.1083	0.2511	0.5827
No log	7.0	175	0.0777	0.275	0.8944	5.1450	0.275	0.1053	0.2840	0.5127
No log	8.0	200	0.0767	0.305	0.8920	5.0689	0.305	0.1177	0.3095	0.4372
No log	9.0	225	0.0756	0.305	0.8893	4.6689	0.305	0.1015	0.3114	0.4279
No log	10.0	250	0.0747	0.325	0.8865	5.1088	0.325	0.1067	0.3257	0.3891
No log	11.0	275	0.0738	0.345	0.8803	4.9911	0.345	0.1297	0.3418	0.3672
No log	12.0	300	0.0729	0.405	0.8748	4.7892	0.405	0.1649	0.3774	0.3276
No log	13.0	325	0.0721	0.395	0.8696	4.6346	0.395	0.1561	0.3723	0.3221
No log	14.0	350	0.0713	0.44	0.8676	3.8051	0.44	0.1780	0.4047	0.2743
No log	15.0	375	0.0705	0.43	0.8612	3.6440	0.4300	0.1893	0.3994	0.2701
No log	16.0	400	0.0698	0.48	0.8569	3.4952	0.48	0.2607	0.4216	0.2640
No log	17.0	425	0.0690	0.54	0.8533	3.2018	0.54	0.3042	0.4771	0.2168
No log	18.0	450	0.0683	0.54	0.8500	2.7042	0.54	0.3161	0.4677	0.2075
No log	19.0	475	0.0679	0.6	0.8469	2.5343	0.6	0.4246	0.5186	0.1865
0.0767	20.0	500	0.0677	0.615	0.8455	2.6485	0.615	0.4407	0.5207	0.1908
0.0767	21.0	525	0.0673	0.625	0.8463	2.2841	0.625	0.4471	0.5369	0.1558
0.0767	22.0	550	0.0667	0.645	0.8398	2.4032	0.645	0.4633	0.5462	0.1500
0.0767	23.0	575	0.0664	0.63	0.8388	2.4376	0.63	0.4600	0.5356	0.1583
0.0767	24.0	600	0.0663	0.645	0.8371	2.3057	0.645	0.4731	0.5443	0.1437
0.0767	25.0	625	0.0662	0.635	0.8386	2.2486	0.635	0.4606	0.5425	0.1515
0.0767	26.0	650	0.0661	0.63	0.8374	2.2367	0.63	0.4543	0.5423	0.1549
0.0767	27.0	675	0.0660	0.64	0.8358	2.1278	0.64	0.4554	0.5486	0.1350
0.0767	28.0	700	0.0660	0.64	0.8360	2.2416	0.64	0.4726	0.5363	0.1429
0.0767	29.0	725	0.0660	0.67	0.8364	2.1574	0.67	0.4990	0.5648	0.1264
0.0767	30.0	750	0.0659	0.665	0.8357	2.2015	0.665	0.5113	0.5645	0.1383
0.0767	31.0	775	0.0658	0.65	0.8347	2.1367	0.65	0.4995	0.5522	0.1461
0.0767	32.0	800	0.0656	0.67	0.8341	2.1025	0.67	0.5110	0.5666	0.1307
0.0767	33.0	825	0.0656	0.645	0.8354	2.0398	0.645	0.5034	0.5442	0.1334
0.0767	34.0	850	0.0656	0.67	0.8346	2.1934	0.67	0.5112	0.5569	0.1299
0.0767	35.0	875	0.0658	0.665	0.8353	2.0671	0.665	0.5255	0.5646	0.1295
0.0767	36.0	900	0.0655	0.665	0.8320	2.0168	0.665	0.5138	0.5680	0.1306
0.0767	37.0	925	0.0655	0.675	0.8315	2.0974	0.675	0.5229	0.5672	0.1333
0.0767	38.0	950	0.0655	0.675	0.8341	2.0624	0.675	0.5457	0.5750	0.1256
0.0767	39.0	975	0.0653	0.69	0.8321	2.0556	0.69	0.5498	0.5856	0.1250
0.0625	40.0	1000	0.0653	0.69	0.8330	1.9627	0.69	0.5812	0.5765	0.1243
0.0625	41.0	1025	0.0653	0.705	0.8335	2.0491	0.705	0.5900	0.5919	0.1155
0.0625	42.0	1050	0.0653	0.705	0.8335	2.0357	0.705	0.5984	0.5945	0.1250
0.0625	43.0	1075	0.0652	0.7	0.8316	2.0326	0.7	0.5957	0.5932	0.1230
0.0625	44.0	1100	0.0653	0.69	0.8323	2.0244	0.69	0.5904	0.5911	0.1252
0.0625	45.0	1125	0.0653	0.68	0.8310	2.0410	0.68	0.5644	0.5699	0.1305
0.0625	46.0	1150	0.0653	0.695	0.8323	2.0288	0.695	0.5944	0.5837	0.1251
0.0625	47.0	1175	0.0652	0.685	0.8312	1.9613	0.685	0.5894	0.5834	0.1244
0.0625	48.0	1200	0.0652	0.685	0.8312	1.9620	0.685	0.5753	0.5728	0.1321
0.0625	49.0	1225	0.0652	0.695	0.8317	1.9706	0.695	0.5962	0.5837	0.1291
0.0625	50.0	1250	0.0651	0.69	0.8314	1.9661	0.69	0.5902	0.5759	0.1315
0.0625	51.0	1275	0.0652	0.68	0.8319	1.9542	0.68	0.5695	0.5704	0.1288
0.0625	52.0	1300	0.0651	0.695	0.8308	1.9577	0.695	0.5834	0.5823	0.1276
0.0625	53.0	1325	0.0652	0.67	0.8315	1.8876	0.67	0.5604	0.5680	0.1326
0.0625	54.0	1350	0.0651	0.68	0.8318	1.8731	0.68	0.5925	0.5644	0.1317
0.0625	55.0	1375	0.0651	0.7	0.8292	1.9448	0.7	0.5856	0.5903	0.1214
0.0625	56.0	1400	0.0652	0.705	0.8310	2.0042	0.705	0.6059	0.5881	0.1195
0.0625	57.0	1425	0.0651	0.685	0.8309	1.9467	0.685	0.5832	0.5734	0.1273
0.0625	58.0	1450	0.0651	0.705	0.8306	1.9480	0.705	0.6064	0.5956	0.1227
0.0625	59.0	1475	0.0651	0.695	0.8302	1.9453	0.695	0.5998	0.5806	0.1310
0.0604	60.0	1500	0.0651	0.68	0.8305	1.8892	0.68	0.5813	0.5643	0.1276
0.0604	61.0	1525	0.0651	0.725	0.8302	1.9304	0.7250	0.6346	0.6022	0.1194
0.0604	62.0	1550	0.0651	0.685	0.8303	1.8831	0.685	0.5773	0.5815	0.1322
0.0604	63.0	1575	0.0650	0.71	0.8299	1.9502	0.7100	0.6140	0.5944	0.1257
0.0604	64.0	1600	0.0651	0.68	0.8296	1.9407	0.68	0.5701	0.5727	0.1337
0.0604	65.0	1625	0.0651	0.695	0.8309	1.9413	0.695	0.5995	0.5884	0.1234
0.0604	66.0	1650	0.0651	0.69	0.8298	1.9474	0.69	0.5865	0.5723	0.1293
0.0604	67.0	1675	0.0650	0.705	0.8298	1.8996	0.705	0.6109	0.5966	0.1258
0.0604	68.0	1700	0.0651	0.7	0.8298	1.9938	0.7	0.6089	0.5895	0.1283
0.0604	69.0	1725	0.0651	0.695	0.8296	1.9273	0.695	0.5923	0.5776	0.1251
0.0604	70.0	1750	0.0651	0.705	0.8297	1.8920	0.705	0.6162	0.5868	0.1323
0.0604	71.0	1775	0.0651	0.7	0.8304	1.9852	0.7	0.6123	0.5878	0.1282
0.0604	72.0	1800	0.0651	0.68	0.8310	1.9399	0.68	0.5963	0.5633	0.1345
0.0604	73.0	1825	0.0650	0.725	0.8302	1.9237	0.7250	0.6266	0.6029	0.1192
0.0604	74.0	1850	0.0651	0.68	0.8306	1.9521	0.68	0.5967	0.5745	0.1342
0.0604	75.0	1875	0.0651	0.695	0.8301	1.9911	0.695	0.6047	0.5841	0.1317
0.0604	76.0	1900	0.0651	0.695	0.8299	1.9333	0.695	0.5935	0.5715	0.1299
0.0604	77.0	1925	0.0651	0.695	0.8298	1.9429	0.695	0.6041	0.5679	0.1293
0.0604	78.0	1950	0.0651	0.695	0.8298	1.9367	0.695	0.6101	0.5792	0.1279
0.0604	79.0	1975	0.0651	0.695	0.8301	1.9934	0.695	0.6095	0.5898	0.1324
0.0596	80.0	2000	0.0651	0.7	0.8297	1.9332	0.7	0.6071	0.5778	0.1271
0.0596	81.0	2025	0.0651	0.685	0.8303	1.9457	0.685	0.5986	0.5807	0.1320
0.0596	82.0	2050	0.0651	0.7	0.8300	1.9337	0.7	0.6072	0.5896	0.1296
0.0596	83.0	2075	0.0651	0.685	0.8298	1.9424	0.685	0.5985	0.5753	0.1319
0.0596	84.0	2100	0.0651	0.7	0.8297	1.9407	0.7	0.6116	0.5847	0.1311
0.0596	85.0	2125	0.0651	0.685	0.8298	1.9364	0.685	0.5983	0.5841	0.1311
0.0596	86.0	2150	0.0651	0.685	0.8299	1.9345	0.685	0.5983	0.5806	0.1318
0.0596	87.0	2175	0.0652	0.685	0.8299	1.9402	0.685	0.5979	0.5778	0.1317
0.0596	88.0	2200	0.0651	0.685	0.8298	1.9385	0.685	0.5983	0.5726	0.1315
0.0596	89.0	2225	0.0652	0.68	0.8296	1.9367	0.68	0.5899	0.5732	0.1314
0.0596	90.0	2250	0.0652	0.68	0.8298	1.9383	0.68	0.5896	0.5782	0.1321
0.0596	91.0	2275	0.0652	0.68	0.8297	1.9408	0.68	0.5896	0.5782	0.1317
0.0596	92.0	2300	0.0652	0.68	0.8299	1.9370	0.68	0.5899	0.5701	0.1320
0.0596	93.0	2325	0.0652	0.68	0.8298	1.9395	0.68	0.5899	0.5754	0.1321
0.0596	94.0	2350	0.0652	0.68	0.8297	1.9392	0.68	0.5899	0.5701	0.1326
0.0596	95.0	2375	0.0652	0.68	0.8297	1.9393	0.68	0.5899	0.5651	0.1320
0.0596	96.0	2400	0.0652	0.68	0.8297	1.9397	0.68	0.5899	0.5701	0.1321
0.0596	97.0	2425	0.0652	0.68	0.8297	1.9400	0.68	0.5899	0.5676	0.1322
0.0596	98.0	2450	0.0652	0.68	0.8297	1.9391	0.68	0.5899	0.5677	0.1320
0.0596	99.0	2475	0.0652	0.68	0.8297	1.9397	0.68	0.5899	0.5701	0.1321
0.0592	100.0	2500	0.0652	0.68	0.8297	1.9403	0.68	0.5899	0.5728	0.1321

Framework versions

Transformers 4.26.1
Pytorch 1.13.1.post200
Datasets 2.9.0
Tokenizers 0.13.2

jordyvl/39-tiny_tobacco3482

39-tiny_tobacco3482

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions