vit-small_tobacco3482_kd_CEKD_t2.5_a0.5

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: 0.4300
Accuracy: 0.83
Brier Loss: 0.2807
Nll: 1.0350
F1 Micro: 0.83
F1 Macro: 0.8295
Ece: 0.2287
Aurc: 0.0560

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	1.6525	0.225	0.8757	5.3231	0.225	0.1387	0.2689	0.6977
No log	2.0	14	1.3106	0.405	0.7470	3.3487	0.405	0.2195	0.2936	0.4032
No log	3.0	21	0.9127	0.585	0.5785	1.8686	0.585	0.5142	0.2974	0.2067
No log	4.0	28	0.7280	0.715	0.4339	1.6780	0.715	0.6761	0.2672	0.1204
No log	5.0	35	0.6523	0.775	0.3676	1.6537	0.775	0.7619	0.2554	0.0929
No log	6.0	42	0.5888	0.785	0.3502	1.3926	0.785	0.7538	0.2277	0.0908
No log	7.0	49	0.6113	0.805	0.3326	1.7118	0.805	0.7903	0.2428	0.0803
No log	8.0	56	0.5404	0.785	0.3178	1.1557	0.785	0.7671	0.2183	0.0716
No log	9.0	63	0.5380	0.82	0.3051	1.3231	0.82	0.8072	0.2168	0.0773
No log	10.0	70	0.6035	0.775	0.3508	1.3888	0.775	0.7682	0.2191	0.0812
No log	11.0	77	0.5473	0.795	0.3202	1.2622	0.795	0.7740	0.2303	0.0626
No log	12.0	84	0.4860	0.825	0.2937	1.3575	0.825	0.8053	0.2392	0.0727
No log	13.0	91	0.5046	0.81	0.3032	1.1857	0.81	0.8086	0.2248	0.0564
No log	14.0	98	0.4745	0.825	0.2870	1.2338	0.825	0.8089	0.2441	0.0518
No log	15.0	105	0.4764	0.81	0.2943	1.0325	0.81	0.8110	0.1935	0.0556
No log	16.0	112	0.4918	0.81	0.3062	1.0551	0.81	0.8015	0.2198	0.0587
No log	17.0	119	0.4757	0.815	0.2970	1.4203	0.815	0.7965	0.2263	0.0850
No log	18.0	126	0.4586	0.825	0.2926	1.0361	0.825	0.8268	0.2279	0.0583
No log	19.0	133	0.4503	0.835	0.2855	1.1476	0.835	0.8301	0.2392	0.0589
No log	20.0	140	0.4780	0.805	0.3105	0.9928	0.805	0.7902	0.1988	0.0775
No log	21.0	147	0.4965	0.8	0.3205	1.1887	0.8000	0.8029	0.2410	0.0702
No log	22.0	154	0.4753	0.815	0.3016	0.9609	0.815	0.8169	0.2163	0.0580
No log	23.0	161	0.4733	0.8	0.3074	1.2566	0.8000	0.8001	0.2162	0.0704
No log	24.0	168	0.4472	0.815	0.2888	1.0352	0.815	0.8187	0.2317	0.0590
No log	25.0	175	0.4434	0.815	0.2854	0.9874	0.815	0.8186	0.2149	0.0554
No log	26.0	182	0.4316	0.82	0.2754	1.0477	0.82	0.8267	0.2195	0.0508
No log	27.0	189	0.4276	0.83	0.2751	1.1016	0.83	0.8336	0.2050	0.0525
No log	28.0	196	0.4329	0.82	0.2795	1.0537	0.82	0.8220	0.2158	0.0611
No log	29.0	203	0.4327	0.82	0.2827	1.1766	0.82	0.8237	0.2024	0.0603
No log	30.0	210	0.4317	0.82	0.2820	1.0331	0.82	0.8219	0.2083	0.0611
No log	31.0	217	0.4316	0.82	0.2803	1.0974	0.82	0.8263	0.1984	0.0575
No log	32.0	224	0.4340	0.82	0.2833	1.0384	0.82	0.8240	0.2202	0.0590
No log	33.0	231	0.4333	0.81	0.2824	1.0355	0.81	0.8160	0.2103	0.0586
No log	34.0	238	0.4309	0.83	0.2817	1.1015	0.83	0.8307	0.2107	0.0577
No log	35.0	245	0.4321	0.82	0.2817	1.0359	0.82	0.8229	0.2147	0.0590
No log	36.0	252	0.4304	0.825	0.2802	1.1016	0.825	0.8257	0.2137	0.0569
No log	37.0	259	0.4303	0.825	0.2811	1.0990	0.825	0.8268	0.2149	0.0581
No log	38.0	266	0.4314	0.825	0.2814	1.1003	0.825	0.8257	0.2163	0.0581
No log	39.0	273	0.4302	0.82	0.2806	1.1007	0.82	0.8226	0.2102	0.0576
No log	40.0	280	0.4307	0.825	0.2809	1.0376	0.825	0.8264	0.2049	0.0573
No log	41.0	287	0.4303	0.82	0.2808	1.0434	0.82	0.8226	0.2096	0.0574
No log	42.0	294	0.4310	0.825	0.2817	1.0376	0.825	0.8268	0.2140	0.0580
No log	43.0	301	0.4310	0.825	0.2813	1.0391	0.825	0.8257	0.2147	0.0580
No log	44.0	308	0.4301	0.825	0.2808	1.0389	0.825	0.8257	0.2064	0.0573
No log	45.0	315	0.4305	0.83	0.2811	1.0419	0.83	0.8307	0.2300	0.0577
No log	46.0	322	0.4303	0.82	0.2808	1.0423	0.82	0.8226	0.2197	0.0582
No log	47.0	329	0.4304	0.825	0.2811	1.0405	0.825	0.8257	0.2240	0.0580
No log	48.0	336	0.4300	0.82	0.2805	1.0407	0.82	0.8226	0.2105	0.0574
No log	49.0	343	0.4307	0.825	0.2812	1.0381	0.825	0.8257	0.2252	0.0577
No log	50.0	350	0.4304	0.82	0.2810	1.0422	0.82	0.8226	0.2353	0.0578
No log	51.0	357	0.4310	0.825	0.2813	1.0382	0.825	0.8264	0.2153	0.0569
No log	52.0	364	0.4309	0.82	0.2814	1.0380	0.82	0.8226	0.2282	0.0574
No log	53.0	371	0.4307	0.825	0.2813	1.0357	0.825	0.8264	0.2250	0.0568
No log	54.0	378	0.4305	0.82	0.2810	1.0366	0.82	0.8226	0.2284	0.0575
No log	55.0	385	0.4304	0.825	0.2811	1.0351	0.825	0.8264	0.2241	0.0566
No log	56.0	392	0.4308	0.825	0.2813	1.0369	0.825	0.8257	0.2414	0.0572
No log	57.0	399	0.4305	0.825	0.2810	1.0356	0.825	0.8257	0.2322	0.0571
No log	58.0	406	0.4302	0.82	0.2808	1.0359	0.82	0.8226	0.2368	0.0569
No log	59.0	413	0.4302	0.82	0.2809	1.0346	0.82	0.8226	0.2271	0.0569
No log	60.0	420	0.4303	0.82	0.2809	1.0357	0.82	0.8226	0.2272	0.0570
No log	61.0	427	0.4304	0.825	0.2810	1.0360	0.825	0.8257	0.2325	0.0569
No log	62.0	434	0.4303	0.825	0.2809	1.0360	0.825	0.8257	0.2321	0.0568
No log	63.0	441	0.4303	0.83	0.2809	1.0356	0.83	0.8295	0.2300	0.0562
No log	64.0	448	0.4304	0.825	0.2810	1.0347	0.825	0.8264	0.2242	0.0564
No log	65.0	455	0.4301	0.83	0.2808	1.0361	0.83	0.8295	0.2384	0.0564
No log	66.0	462	0.4303	0.83	0.2810	1.0359	0.83	0.8295	0.2293	0.0563
No log	67.0	469	0.4302	0.83	0.2809	1.0360	0.83	0.8295	0.2386	0.0564
No log	68.0	476	0.4304	0.83	0.2810	1.0360	0.83	0.8295	0.2384	0.0563
No log	69.0	483	0.4305	0.83	0.2812	1.0355	0.83	0.8295	0.2295	0.0564
No log	70.0	490	0.4302	0.825	0.2808	1.0354	0.825	0.8264	0.2239	0.0561
No log	71.0	497	0.4305	0.83	0.2812	1.0352	0.83	0.8295	0.2296	0.0564
0.1776	72.0	504	0.4303	0.83	0.2808	1.0356	0.83	0.8295	0.2287	0.0561
0.1776	73.0	511	0.4301	0.825	0.2807	1.0351	0.825	0.8264	0.2348	0.0563
0.1776	74.0	518	0.4304	0.83	0.2811	1.0353	0.83	0.8295	0.2195	0.0562
0.1776	75.0	525	0.4301	0.825	0.2808	1.0355	0.825	0.8257	0.2320	0.0568
0.1776	76.0	532	0.4302	0.83	0.2808	1.0348	0.83	0.8295	0.2289	0.0561
0.1776	77.0	539	0.4301	0.83	0.2808	1.0355	0.83	0.8295	0.2300	0.0562
0.1776	78.0	546	0.4301	0.83	0.2808	1.0354	0.83	0.8295	0.2394	0.0563
0.1776	79.0	553	0.4302	0.83	0.2809	1.0346	0.83	0.8295	0.2287	0.0560
0.1776	80.0	560	0.4302	0.83	0.2809	1.0353	0.83	0.8295	0.2299	0.0563
0.1776	81.0	567	0.4302	0.83	0.2809	1.0350	0.83	0.8295	0.2299	0.0563
0.1776	82.0	574	0.4302	0.83	0.2808	1.0354	0.83	0.8295	0.2298	0.0560
0.1776	83.0	581	0.4302	0.83	0.2809	1.0350	0.83	0.8295	0.2299	0.0561
0.1776	84.0	588	0.4299	0.83	0.2807	1.0352	0.83	0.8295	0.2287	0.0561
0.1776	85.0	595	0.4301	0.83	0.2808	1.0349	0.83	0.8295	0.2296	0.0562
0.1776	86.0	602	0.4301	0.83	0.2808	1.0351	0.83	0.8295	0.2287	0.0562
0.1776	87.0	609	0.4300	0.83	0.2807	1.0351	0.83	0.8295	0.2297	0.0561
0.1776	88.0	616	0.4300	0.83	0.2807	1.0349	0.83	0.8295	0.2287	0.0562
0.1776	89.0	623	0.4300	0.83	0.2807	1.0353	0.83	0.8295	0.2296	0.0560
0.1776	90.0	630	0.4300	0.83	0.2807	1.0349	0.83	0.8295	0.2297	0.0559
0.1776	91.0	637	0.4300	0.83	0.2807	1.0352	0.83	0.8295	0.2296	0.0562
0.1776	92.0	644	0.4300	0.83	0.2807	1.0351	0.83	0.8295	0.2287	0.0561
0.1776	93.0	651	0.4300	0.83	0.2807	1.0351	0.83	0.8295	0.2297	0.0562
0.1776	94.0	658	0.4300	0.83	0.2807	1.0349	0.83	0.8295	0.2297	0.0560
0.1776	95.0	665	0.4300	0.83	0.2807	1.0350	0.83	0.8295	0.2297	0.0562
0.1776	96.0	672	0.4300	0.83	0.2807	1.0349	0.83	0.8295	0.2296	0.0561
0.1776	97.0	679	0.4300	0.83	0.2807	1.0350	0.83	0.8295	0.2296	0.0560
0.1776	98.0	686	0.4300	0.83	0.2807	1.0350	0.83	0.8295	0.2296	0.0560
0.1776	99.0	693	0.4300	0.83	0.2807	1.0350	0.83	0.8295	0.2287	0.0560
0.1776	100.0	700	0.4300	0.83	0.2807	1.0350	0.83	0.8295	0.2287	0.0560

Framework versions

Transformers 4.26.1
Pytorch 1.13.1.post200
Datasets 2.9.0
Tokenizers 0.13.2

vit-small_tobacco3482_kd_CEKD_t2.5_a0.5

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

NSDT 3DConvert

UnrealSynth

DreamTexture.js