39-tiny_tobacco3482_kd_CEKD_t2.5_a0.5

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: 0.5285
Accuracy: 0.81
Brier Loss: 0.3907
Nll: 0.9159
F1 Micro: 0.81
F1 Macro: 0.7907
Ece: 0.3421
Aurc: 0.0542

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	1.8936	0.11	1.0097	8.5078	0.11	0.0902	0.3251	0.8953
No log	2.0	14	1.2597	0.16	0.8753	5.5353	0.16	0.1308	0.2539	0.7902
No log	3.0	21	1.0230	0.355	0.7833	5.1396	0.3550	0.2872	0.2810	0.4521
No log	4.0	28	0.8743	0.545	0.6497	3.0412	0.545	0.4328	0.3224	0.2775
No log	5.0	35	0.8020	0.625	0.5958	2.6099	0.625	0.5465	0.3186	0.2136
No log	6.0	42	0.7221	0.675	0.5300	2.3085	0.675	0.5632	0.3257	0.1562
No log	7.0	49	0.6964	0.68	0.4843	1.9033	0.68	0.5761	0.3039	0.1453
No log	8.0	56	0.6729	0.72	0.4598	1.8200	0.72	0.6195	0.3089	0.1170
No log	9.0	63	0.6470	0.77	0.4318	1.5607	0.7700	0.7058	0.3518	0.0897
No log	10.0	70	0.5889	0.795	0.4019	1.1238	0.795	0.7546	0.3324	0.0675
No log	11.0	77	0.5829	0.795	0.4013	1.0267	0.795	0.7667	0.3142	0.0728
No log	12.0	84	0.5763	0.785	0.3923	1.2697	0.785	0.7655	0.3286	0.0751
No log	13.0	91	0.5854	0.765	0.3934	1.4915	0.765	0.7291	0.2936	0.0806
No log	14.0	98	0.5779	0.795	0.3983	1.2207	0.795	0.7409	0.3141	0.0681
No log	15.0	105	0.5564	0.795	0.3752	1.1974	0.795	0.7687	0.3201	0.0626
No log	16.0	112	0.5599	0.815	0.3945	1.0987	0.815	0.7827	0.3233	0.0618
No log	17.0	119	0.5748	0.77	0.4001	1.2395	0.7700	0.7497	0.3136	0.0866
No log	18.0	126	0.5611	0.79	0.4028	1.3279	0.79	0.7738	0.3127	0.0680
No log	19.0	133	0.5514	0.805	0.4063	0.8598	0.805	0.7873	0.3656	0.0575
No log	20.0	140	0.5566	0.81	0.4028	0.9944	0.81	0.7943	0.3449	0.0676
No log	21.0	147	0.5489	0.81	0.3879	1.1351	0.81	0.7966	0.3432	0.0682
No log	22.0	154	0.5586	0.82	0.4091	1.1107	0.82	0.7894	0.3526	0.0580
No log	23.0	161	0.5593	0.795	0.4131	1.1693	0.795	0.7765	0.3483	0.0641
No log	24.0	168	0.5493	0.79	0.3962	1.2363	0.79	0.7740	0.3494	0.0646
No log	25.0	175	0.5489	0.8	0.3930	1.0310	0.8000	0.7638	0.3342	0.0614
No log	26.0	182	0.5492	0.79	0.3944	1.3201	0.79	0.7670	0.3096	0.0667
No log	27.0	189	0.5441	0.805	0.4002	1.1304	0.805	0.7886	0.3528	0.0600
No log	28.0	196	0.5397	0.815	0.3960	1.1210	0.815	0.7902	0.3630	0.0544
No log	29.0	203	0.5418	0.785	0.3977	0.9580	0.785	0.7575	0.3536	0.0646
No log	30.0	210	0.5374	0.815	0.3931	1.0186	0.815	0.7855	0.3422	0.0604
No log	31.0	217	0.5405	0.815	0.3983	0.8948	0.815	0.7980	0.3671	0.0531
No log	32.0	224	0.5394	0.805	0.3998	1.0680	0.805	0.7841	0.3695	0.0568
No log	33.0	231	0.5296	0.81	0.3868	1.1222	0.81	0.7891	0.3530	0.0545
No log	34.0	238	0.5338	0.81	0.3952	1.1333	0.81	0.7825	0.3453	0.0559
No log	35.0	245	0.5339	0.805	0.3941	0.8600	0.805	0.7905	0.3552	0.0554
No log	36.0	252	0.5332	0.81	0.3918	0.9018	0.81	0.7996	0.3669	0.0527
No log	37.0	259	0.5336	0.79	0.3907	0.7768	0.79	0.7612	0.3374	0.0611
No log	38.0	266	0.5327	0.805	0.3906	0.9987	0.805	0.7750	0.3430	0.0564
No log	39.0	273	0.5342	0.805	0.3898	1.1024	0.805	0.7837	0.3295	0.0563
No log	40.0	280	0.5310	0.81	0.3906	0.8426	0.81	0.7820	0.3513	0.0556
No log	41.0	287	0.5327	0.81	0.3950	1.0952	0.81	0.7927	0.3418	0.0570
No log	42.0	294	0.5305	0.82	0.3961	0.7830	0.82	0.8011	0.3501	0.0545
No log	43.0	301	0.5308	0.81	0.3926	0.9752	0.81	0.7907	0.3534	0.0573
No log	44.0	308	0.5287	0.81	0.3898	0.9838	0.81	0.7904	0.3454	0.0570
No log	45.0	315	0.5270	0.815	0.3890	0.8682	0.815	0.8004	0.3499	0.0543
No log	46.0	322	0.5272	0.81	0.3884	0.9784	0.81	0.7827	0.3415	0.0541
No log	47.0	329	0.5306	0.805	0.3900	1.1153	0.805	0.7800	0.3388	0.0571
No log	48.0	336	0.5288	0.82	0.3915	0.9916	0.82	0.7912	0.3519	0.0527
No log	49.0	343	0.5274	0.81	0.3886	0.8415	0.81	0.7855	0.3524	0.0550
No log	50.0	350	0.5264	0.81	0.3868	0.9713	0.81	0.7907	0.3408	0.0559
No log	51.0	357	0.5295	0.815	0.3916	1.0340	0.815	0.7933	0.3683	0.0536
No log	52.0	364	0.5294	0.81	0.3920	0.9178	0.81	0.7854	0.3499	0.0563
No log	53.0	371	0.5283	0.81	0.3912	0.8517	0.81	0.7907	0.3648	0.0540
No log	54.0	378	0.5301	0.815	0.3927	0.9279	0.815	0.7933	0.3579	0.0558
No log	55.0	385	0.5275	0.805	0.3888	0.9225	0.805	0.7800	0.3406	0.0553
No log	56.0	392	0.5284	0.815	0.3903	0.9064	0.815	0.7933	0.3463	0.0551
No log	57.0	399	0.5261	0.81	0.3872	0.9072	0.81	0.7907	0.3527	0.0551
No log	58.0	406	0.5278	0.815	0.3900	0.8469	0.815	0.7966	0.3622	0.0526
No log	59.0	413	0.5280	0.81	0.3900	0.9220	0.81	0.7907	0.3467	0.0551
No log	60.0	420	0.5296	0.81	0.3932	0.9166	0.81	0.7907	0.3620	0.0555
No log	61.0	427	0.5288	0.815	0.3925	0.8647	0.815	0.7966	0.3491	0.0529
No log	62.0	434	0.5288	0.81	0.3909	0.9205	0.81	0.7907	0.3482	0.0552
No log	63.0	441	0.5274	0.81	0.3889	0.9143	0.81	0.7907	0.3457	0.0541
No log	64.0	448	0.5283	0.81	0.3905	0.9141	0.81	0.7907	0.3578	0.0549
No log	65.0	455	0.5283	0.81	0.3907	0.9177	0.81	0.7907	0.3536	0.0548
No log	66.0	462	0.5289	0.81	0.3912	0.9179	0.81	0.7907	0.3502	0.0550
No log	67.0	469	0.5282	0.81	0.3903	0.9134	0.81	0.7907	0.3511	0.0547
No log	68.0	476	0.5279	0.81	0.3901	0.9105	0.81	0.7907	0.3473	0.0541
No log	69.0	483	0.5283	0.81	0.3907	0.9128	0.81	0.7907	0.3558	0.0539
No log	70.0	490	0.5283	0.81	0.3904	0.9191	0.81	0.7907	0.3414	0.0543
No log	71.0	497	0.5284	0.81	0.3905	0.9183	0.81	0.7907	0.3478	0.0546
0.3962	72.0	504	0.5285	0.81	0.3909	0.9151	0.81	0.7907	0.3415	0.0545
0.3962	73.0	511	0.5283	0.81	0.3906	0.9144	0.81	0.7907	0.3499	0.0542
0.3962	74.0	518	0.5282	0.81	0.3903	0.9146	0.81	0.7907	0.3411	0.0541
0.3962	75.0	525	0.5284	0.81	0.3909	0.9159	0.81	0.7907	0.3571	0.0542
0.3962	76.0	532	0.5284	0.81	0.3906	0.9155	0.81	0.7907	0.3361	0.0543
0.3962	77.0	539	0.5283	0.81	0.3906	0.9159	0.81	0.7907	0.3480	0.0541
0.3962	78.0	546	0.5282	0.81	0.3905	0.9120	0.81	0.7907	0.3413	0.0540
0.3962	79.0	553	0.5283	0.81	0.3905	0.9162	0.81	0.7907	0.3412	0.0542
0.3962	80.0	560	0.5285	0.81	0.3907	0.9189	0.81	0.7907	0.3361	0.0543
0.3962	81.0	567	0.5285	0.81	0.3907	0.9162	0.81	0.7907	0.3470	0.0541
0.3962	82.0	574	0.5283	0.81	0.3904	0.9144	0.81	0.7907	0.3411	0.0540
0.3962	83.0	581	0.5284	0.81	0.3906	0.9153	0.81	0.7907	0.3361	0.0542
0.3962	84.0	588	0.5284	0.81	0.3907	0.9151	0.81	0.7907	0.3419	0.0542
0.3962	85.0	595	0.5283	0.81	0.3905	0.9143	0.81	0.7907	0.3362	0.0541
0.3962	86.0	602	0.5285	0.81	0.3908	0.9152	0.81	0.7907	0.3418	0.0540
0.3962	87.0	609	0.5284	0.81	0.3907	0.9156	0.81	0.7907	0.3365	0.0543
0.3962	88.0	616	0.5285	0.81	0.3907	0.9155	0.81	0.7907	0.3419	0.0541
0.3962	89.0	623	0.5284	0.81	0.3906	0.9154	0.81	0.7907	0.3360	0.0541
0.3962	90.0	630	0.5285	0.81	0.3907	0.9168	0.81	0.7907	0.3418	0.0543
0.3962	91.0	637	0.5285	0.81	0.3907	0.9160	0.81	0.7907	0.3420	0.0543
0.3962	92.0	644	0.5285	0.81	0.3908	0.9164	0.81	0.7907	0.3421	0.0541
0.3962	93.0	651	0.5285	0.81	0.3907	0.9164	0.81	0.7907	0.3473	0.0542
0.3962	94.0	658	0.5285	0.81	0.3907	0.9164	0.81	0.7907	0.3420	0.0542
0.3962	95.0	665	0.5285	0.81	0.3907	0.9161	0.81	0.7907	0.3473	0.0541
0.3962	96.0	672	0.5285	0.81	0.3907	0.9157	0.81	0.7907	0.3421	0.0542
0.3962	97.0	679	0.5285	0.81	0.3907	0.9154	0.81	0.7907	0.3363	0.0542
0.3962	98.0	686	0.5285	0.81	0.3907	0.9164	0.81	0.7907	0.3420	0.0542
0.3962	99.0	693	0.5285	0.81	0.3907	0.9162	0.81	0.7907	0.3420	0.0542
0.3962	100.0	700	0.5285	0.81	0.3907	0.9159	0.81	0.7907	0.3421	0.0542

Framework versions

Transformers 4.26.1
Pytorch 1.13.1.post200
Datasets 2.9.0
Tokenizers 0.13.2

39-tiny_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