vit-small_tobacco3482_kd_CEKD_t1.5_a0.7

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: 0.4797
Accuracy: 0.835
Brier Loss: 0.2522
Nll: 0.8627
F1 Micro: 0.835
F1 Macro: 0.8222
Ece: 0.1830
Aurc: 0.0434

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	1.9341	0.215	0.8749	5.3238	0.2150	0.1264	0.2642	0.6914
No log	2.0	14	1.5320	0.405	0.7410	3.5078	0.405	0.2276	0.2957	0.4015
No log	3.0	21	1.0532	0.635	0.5629	2.0153	0.635	0.5844	0.3037	0.2006
No log	4.0	28	0.7915	0.715	0.4093	1.6974	0.715	0.6762	0.2420	0.1131
No log	5.0	35	0.8024	0.745	0.3869	1.7109	0.745	0.7548	0.2160	0.1006
No log	6.0	42	0.7162	0.765	0.3351	1.8105	0.765	0.7599	0.2216	0.0874
No log	7.0	49	0.6966	0.785	0.3304	1.5292	0.785	0.7682	0.2058	0.0979
No log	8.0	56	0.6317	0.805	0.2995	1.3486	0.805	0.7887	0.2266	0.0721
No log	9.0	63	0.6903	0.805	0.3304	1.5866	0.805	0.7971	0.2371	0.0995
No log	10.0	70	0.6223	0.805	0.2940	1.3478	0.805	0.8114	0.2281	0.0697
No log	11.0	77	0.6350	0.795	0.3145	1.3386	0.795	0.7730	0.2063	0.0962
No log	12.0	84	0.5570	0.835	0.2666	1.2662	0.835	0.8181	0.1951	0.0553
No log	13.0	91	0.5610	0.81	0.2858	1.2619	0.81	0.8002	0.1884	0.0626
No log	14.0	98	0.5843	0.8	0.2961	1.0782	0.8000	0.8083	0.1993	0.0683
No log	15.0	105	0.5918	0.78	0.2965	1.1207	0.78	0.7861	0.1895	0.0634
No log	16.0	112	0.5541	0.84	0.2765	1.3189	0.8400	0.8455	0.1969	0.0597
No log	17.0	119	0.5037	0.835	0.2568	0.9024	0.835	0.8248	0.2083	0.0499
No log	18.0	126	0.5050	0.85	0.2563	1.0032	0.85	0.8441	0.2147	0.0580
No log	19.0	133	0.5430	0.815	0.2779	1.1046	0.815	0.8044	0.1906	0.0562
No log	20.0	140	0.5276	0.84	0.2743	0.9964	0.8400	0.8144	0.2104	0.0597
No log	21.0	147	0.5155	0.835	0.2686	0.9556	0.835	0.8210	0.1962	0.0572
No log	22.0	154	0.4937	0.835	0.2581	1.0079	0.835	0.8172	0.1975	0.0479
No log	23.0	161	0.4931	0.845	0.2533	1.0021	0.845	0.8270	0.1884	0.0503
No log	24.0	168	0.4869	0.83	0.2554	0.9660	0.83	0.8084	0.1945	0.0481
No log	25.0	175	0.4843	0.845	0.2512	0.9979	0.845	0.8316	0.1746	0.0466
No log	26.0	182	0.4866	0.835	0.2531	0.9006	0.835	0.8188	0.1833	0.0472
No log	27.0	189	0.4882	0.825	0.2562	0.8929	0.825	0.8043	0.2023	0.0469
No log	28.0	196	0.4814	0.82	0.2494	0.9122	0.82	0.8060	0.1773	0.0451
No log	29.0	203	0.4749	0.835	0.2501	0.8770	0.835	0.8252	0.1688	0.0442
No log	30.0	210	0.4761	0.84	0.2490	0.8848	0.8400	0.8250	0.2068	0.0443
No log	31.0	217	0.4787	0.845	0.2508	0.8754	0.845	0.8309	0.1635	0.0438
No log	32.0	224	0.4791	0.835	0.2521	0.8711	0.835	0.8224	0.1876	0.0446
No log	33.0	231	0.4779	0.84	0.2509	0.8650	0.8400	0.8252	0.1813	0.0436
No log	34.0	238	0.4774	0.84	0.2513	0.8662	0.8400	0.8252	0.1919	0.0441
No log	35.0	245	0.4760	0.835	0.2502	0.8636	0.835	0.8224	0.1840	0.0434
No log	36.0	252	0.4784	0.84	0.2509	0.8688	0.8400	0.8281	0.1691	0.0437
No log	37.0	259	0.4771	0.835	0.2507	0.8670	0.835	0.8224	0.1936	0.0440
No log	38.0	266	0.4764	0.835	0.2499	0.8614	0.835	0.8224	0.1830	0.0434
No log	39.0	273	0.4769	0.835	0.2503	0.8651	0.835	0.8224	0.2001	0.0438
No log	40.0	280	0.4777	0.84	0.2514	0.8608	0.8400	0.8281	0.1832	0.0435
No log	41.0	287	0.4777	0.835	0.2504	0.8650	0.835	0.8224	0.1953	0.0437
No log	42.0	294	0.4779	0.835	0.2511	0.8629	0.835	0.8224	0.1944	0.0440
No log	43.0	301	0.4790	0.835	0.2519	0.8631	0.835	0.8222	0.1808	0.0439
No log	44.0	308	0.4777	0.835	0.2509	0.8604	0.835	0.8222	0.1886	0.0435
No log	45.0	315	0.4787	0.835	0.2517	0.8620	0.835	0.8222	0.1940	0.0437
No log	46.0	322	0.4774	0.84	0.2509	0.8614	0.8400	0.8281	0.1779	0.0433
No log	47.0	329	0.4785	0.835	0.2517	0.8609	0.835	0.8222	0.1811	0.0438
No log	48.0	336	0.4792	0.835	0.2521	0.8611	0.835	0.8222	0.1849	0.0436
No log	49.0	343	0.4771	0.84	0.2509	0.8623	0.8400	0.8281	0.1908	0.0430
No log	50.0	350	0.4793	0.835	0.2520	0.8633	0.835	0.8222	0.1900	0.0435
No log	51.0	357	0.4786	0.83	0.2517	0.8654	0.83	0.8159	0.1684	0.0437
No log	52.0	364	0.4792	0.83	0.2521	0.8625	0.83	0.8166	0.1915	0.0430
No log	53.0	371	0.4785	0.835	0.2513	0.8652	0.835	0.8222	0.1853	0.0434
No log	54.0	378	0.4798	0.835	0.2523	0.8652	0.835	0.8222	0.1767	0.0437
No log	55.0	385	0.4791	0.835	0.2519	0.8637	0.835	0.8222	0.1891	0.0435
No log	56.0	392	0.4790	0.835	0.2519	0.8614	0.835	0.8222	0.1749	0.0429
No log	57.0	399	0.4782	0.835	0.2513	0.8625	0.835	0.8222	0.1909	0.0433
No log	58.0	406	0.4794	0.835	0.2521	0.8602	0.835	0.8222	0.1758	0.0435
No log	59.0	413	0.4790	0.835	0.2517	0.8617	0.835	0.8222	0.1754	0.0432
No log	60.0	420	0.4791	0.835	0.2520	0.8614	0.835	0.8222	0.1830	0.0430
No log	61.0	427	0.4789	0.835	0.2518	0.8612	0.835	0.8222	0.1870	0.0432
No log	62.0	434	0.4792	0.835	0.2520	0.8620	0.835	0.8222	0.1902	0.0433
No log	63.0	441	0.4789	0.835	0.2518	0.8619	0.835	0.8222	0.1997	0.0431
No log	64.0	448	0.4797	0.835	0.2523	0.8607	0.835	0.8222	0.1833	0.0434
No log	65.0	455	0.4797	0.835	0.2522	0.8624	0.835	0.8222	0.1922	0.0434
No log	66.0	462	0.4791	0.835	0.2519	0.8620	0.835	0.8222	0.1894	0.0430
No log	67.0	469	0.4792	0.835	0.2520	0.8612	0.835	0.8222	0.1885	0.0433
No log	68.0	476	0.4796	0.835	0.2522	0.8627	0.835	0.8222	0.1918	0.0433
No log	69.0	483	0.4793	0.835	0.2521	0.8628	0.835	0.8222	0.1828	0.0433
No log	70.0	490	0.4792	0.835	0.2519	0.8622	0.835	0.8222	0.1918	0.0432
No log	71.0	497	0.4797	0.835	0.2523	0.8615	0.835	0.8222	0.1836	0.0434
0.194	72.0	504	0.4797	0.835	0.2522	0.8618	0.835	0.8222	0.1842	0.0433
0.194	73.0	511	0.4794	0.835	0.2521	0.8624	0.835	0.8222	0.1914	0.0432
0.194	74.0	518	0.4794	0.835	0.2521	0.8617	0.835	0.8222	0.1915	0.0431
0.194	75.0	525	0.4796	0.835	0.2522	0.8623	0.835	0.8222	0.1917	0.0434
0.194	76.0	532	0.4795	0.835	0.2520	0.8622	0.835	0.8222	0.1985	0.0433
0.194	77.0	539	0.4795	0.835	0.2520	0.8623	0.835	0.8222	0.1985	0.0432
0.194	78.0	546	0.4795	0.835	0.2522	0.8621	0.835	0.8222	0.1981	0.0432
0.194	79.0	553	0.4798	0.835	0.2522	0.8626	0.835	0.8222	0.1909	0.0433
0.194	80.0	560	0.4796	0.835	0.2521	0.8630	0.835	0.8222	0.1984	0.0433
0.194	81.0	567	0.4797	0.835	0.2522	0.8619	0.835	0.8222	0.1902	0.0434
0.194	82.0	574	0.4797	0.835	0.2522	0.8631	0.835	0.8222	0.1913	0.0433
0.194	83.0	581	0.4797	0.835	0.2522	0.8627	0.835	0.8222	0.1909	0.0433
0.194	84.0	588	0.4797	0.835	0.2522	0.8623	0.835	0.8222	0.1909	0.0433
0.194	85.0	595	0.4797	0.835	0.2522	0.8624	0.835	0.8222	0.1909	0.0434
0.194	86.0	602	0.4796	0.835	0.2522	0.8623	0.835	0.8222	0.1830	0.0433
0.194	87.0	609	0.4797	0.835	0.2522	0.8629	0.835	0.8222	0.1909	0.0434
0.194	88.0	616	0.4797	0.835	0.2521	0.8634	0.835	0.8222	0.1830	0.0433
0.194	89.0	623	0.4797	0.835	0.2522	0.8627	0.835	0.8222	0.1910	0.0434
0.194	90.0	630	0.4798	0.835	0.2523	0.8627	0.835	0.8222	0.1909	0.0434
0.194	91.0	637	0.4797	0.835	0.2522	0.8625	0.835	0.8222	0.1909	0.0434
0.194	92.0	644	0.4797	0.835	0.2522	0.8630	0.835	0.8222	0.1830	0.0434
0.194	93.0	651	0.4798	0.835	0.2522	0.8629	0.835	0.8222	0.1910	0.0434
0.194	94.0	658	0.4797	0.835	0.2522	0.8628	0.835	0.8222	0.1910	0.0434
0.194	95.0	665	0.4797	0.835	0.2522	0.8627	0.835	0.8222	0.1910	0.0434
0.194	96.0	672	0.4798	0.835	0.2522	0.8627	0.835	0.8222	0.1834	0.0435
0.194	97.0	679	0.4797	0.835	0.2522	0.8628	0.835	0.8222	0.1830	0.0434
0.194	98.0	686	0.4797	0.835	0.2522	0.8628	0.835	0.8222	0.1830	0.0434
0.194	99.0	693	0.4797	0.835	0.2522	0.8628	0.835	0.8222	0.1830	0.0434
0.194	100.0	700	0.4797	0.835	0.2522	0.8627	0.835	0.8222	0.1830	0.0434

Framework versions

Transformers 4.26.1
Pytorch 1.13.1.post200
Datasets 2.9.0
Tokenizers 0.13.2

vit-small_tobacco3482_kd_CEKD_t1.5_a0.7

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

NSDT 3DConvert

UnrealSynth

DreamTexture.js