vit-small_tobacco3482_kd_CEKD_t2.5_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: 0.5012
Accuracy: 0.845
Brier Loss: 0.2630
Nll: 0.9559
F1 Micro: 0.845
F1 Macro: 0.8409
Ece: 0.2081
Aurc: 0.0487

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	1.8874	0.22	0.8753	5.3264	0.22	0.1297	0.2491	0.6946
No log	2.0	14	1.5114	0.405	0.7436	3.4265	0.405	0.2289	0.2861	0.4056
No log	3.0	21	1.0555	0.61	0.5700	1.9393	0.61	0.5508	0.3108	0.2025
No log	4.0	28	0.8188	0.695	0.4182	1.7503	0.695	0.6308	0.2374	0.1153
No log	5.0	35	0.8072	0.74	0.3863	1.7844	0.74	0.7491	0.2368	0.1059
No log	6.0	42	0.7259	0.745	0.3520	1.6780	0.745	0.7403	0.2147	0.0935
No log	7.0	49	0.6764	0.805	0.3188	1.5699	0.805	0.7739	0.2217	0.0875
No log	8.0	56	0.6437	0.805	0.3103	1.2385	0.805	0.7880	0.2198	0.0667
No log	9.0	63	0.6173	0.79	0.3033	1.3362	0.79	0.7910	0.2158	0.0688
No log	10.0	70	0.5831	0.83	0.2846	1.1457	0.83	0.8211	0.2293	0.0530
No log	11.0	77	0.6568	0.775	0.3281	1.4185	0.775	0.7253	0.2128	0.0708
No log	12.0	84	0.5838	0.815	0.2859	1.4799	0.815	0.8079	0.2191	0.0518
No log	13.0	91	0.6773	0.75	0.3546	1.1980	0.75	0.7125	0.1970	0.0977
No log	14.0	98	0.5809	0.82	0.2976	1.1146	0.82	0.8166	0.2240	0.0850
No log	15.0	105	0.5589	0.8	0.2913	1.0169	0.8000	0.7942	0.2018	0.0655
No log	16.0	112	0.5495	0.835	0.2857	1.1661	0.835	0.8217	0.1943	0.0800
No log	17.0	119	0.5280	0.83	0.2697	1.0304	0.83	0.8314	0.2216	0.0511
No log	18.0	126	0.5174	0.85	0.2583	1.1388	0.85	0.8392	0.1992	0.0540
No log	19.0	133	0.5271	0.81	0.2838	0.9360	0.81	0.8129	0.1742	0.0622
No log	20.0	140	0.5098	0.845	0.2604	1.1335	0.845	0.8353	0.2040	0.0599
No log	21.0	147	0.5219	0.82	0.2725	0.9910	0.82	0.8145	0.1831	0.0497
No log	22.0	154	0.5195	0.835	0.2706	1.0873	0.835	0.8390	0.1908	0.0515
No log	23.0	161	0.5122	0.835	0.2666	1.0898	0.835	0.8399	0.1998	0.0487
No log	24.0	168	0.5158	0.825	0.2723	1.0534	0.825	0.8210	0.1811	0.0507
No log	25.0	175	0.5059	0.825	0.2654	0.9966	0.825	0.8212	0.1910	0.0487
No log	26.0	182	0.5033	0.825	0.2648	0.9836	0.825	0.8212	0.1768	0.0516
No log	27.0	189	0.5114	0.835	0.2703	0.9847	0.835	0.8353	0.2040	0.0512
No log	28.0	196	0.5047	0.84	0.2654	0.9774	0.8400	0.8359	0.1868	0.0494
No log	29.0	203	0.5027	0.84	0.2656	0.9674	0.8400	0.8359	0.1820	0.0502
No log	30.0	210	0.5035	0.835	0.2660	0.9606	0.835	0.8296	0.1781	0.0494
No log	31.0	217	0.5010	0.835	0.2642	0.9628	0.835	0.8296	0.1893	0.0487
No log	32.0	224	0.5032	0.835	0.2652	0.9705	0.835	0.8296	0.1913	0.0494
No log	33.0	231	0.5052	0.825	0.2664	0.9657	0.825	0.8231	0.1882	0.0503
No log	34.0	238	0.5047	0.825	0.2667	0.9605	0.825	0.8168	0.1938	0.0508
No log	35.0	245	0.5019	0.835	0.2642	0.9596	0.835	0.8296	0.1846	0.0491
No log	36.0	252	0.5035	0.835	0.2648	0.9646	0.835	0.8296	0.2064	0.0492
No log	37.0	259	0.5020	0.835	0.2645	0.9589	0.835	0.8296	0.2036	0.0491
No log	38.0	266	0.5023	0.83	0.2642	0.9595	0.83	0.8262	0.1798	0.0495
No log	39.0	273	0.5023	0.835	0.2643	0.9615	0.835	0.8296	0.1921	0.0491
No log	40.0	280	0.5024	0.835	0.2645	0.9589	0.835	0.8292	0.1813	0.0499
No log	41.0	287	0.5018	0.835	0.2638	0.9583	0.835	0.8296	0.1743	0.0492
No log	42.0	294	0.5018	0.83	0.2640	0.9592	0.83	0.8266	0.1839	0.0498
No log	43.0	301	0.5033	0.84	0.2650	0.9588	0.8400	0.8322	0.1914	0.0497
No log	44.0	308	0.5009	0.83	0.2632	0.9603	0.83	0.8266	0.1652	0.0495
No log	45.0	315	0.5049	0.835	0.2659	0.9587	0.835	0.8344	0.2130	0.0502
No log	46.0	322	0.5018	0.835	0.2637	0.9592	0.835	0.8344	0.1685	0.0496
No log	47.0	329	0.5009	0.835	0.2632	0.9578	0.835	0.8296	0.1971	0.0491
No log	48.0	336	0.5022	0.835	0.2641	0.9574	0.835	0.8288	0.2050	0.0495
No log	49.0	343	0.5017	0.835	0.2635	0.9586	0.835	0.8349	0.2025	0.0493
No log	50.0	350	0.5022	0.84	0.2640	0.9572	0.8400	0.8322	0.1926	0.0493
No log	51.0	357	0.5022	0.835	0.2638	0.9591	0.835	0.8288	0.1948	0.0494
No log	52.0	364	0.5025	0.835	0.2640	0.9573	0.835	0.8344	0.2119	0.0497
No log	53.0	371	0.5022	0.835	0.2638	0.9575	0.835	0.8292	0.2176	0.0497
No log	54.0	378	0.5020	0.83	0.2637	0.9575	0.83	0.8257	0.1935	0.0498
No log	55.0	385	0.5024	0.835	0.2640	0.9572	0.835	0.8288	0.2040	0.0491
No log	56.0	392	0.5023	0.835	0.2639	0.9581	0.835	0.8288	0.2123	0.0492
No log	57.0	399	0.5018	0.84	0.2635	0.9575	0.8400	0.8322	0.2086	0.0487
No log	58.0	406	0.5023	0.835	0.2639	0.9568	0.835	0.8292	0.2094	0.0496
No log	59.0	413	0.5016	0.83	0.2633	0.9563	0.83	0.8257	0.1930	0.0493
No log	60.0	420	0.5015	0.84	0.2633	0.9565	0.8400	0.8322	0.2004	0.0488
No log	61.0	427	0.5017	0.84	0.2635	0.9559	0.8400	0.8322	0.2004	0.0491
No log	62.0	434	0.5018	0.83	0.2635	0.9563	0.83	0.8257	0.1994	0.0497
No log	63.0	441	0.5020	0.835	0.2636	0.9572	0.835	0.8288	0.2040	0.0490
No log	64.0	448	0.5020	0.835	0.2636	0.9565	0.835	0.8288	0.2036	0.0490
No log	65.0	455	0.5017	0.835	0.2634	0.9566	0.835	0.8288	0.1959	0.0490
No log	66.0	462	0.5018	0.84	0.2635	0.9561	0.8400	0.8322	0.2012	0.0491
No log	67.0	469	0.5016	0.84	0.2633	0.9566	0.8400	0.8322	0.1946	0.0489
No log	68.0	476	0.5016	0.84	0.2633	0.9565	0.8400	0.8322	0.1946	0.0488
No log	69.0	483	0.5018	0.835	0.2634	0.9567	0.835	0.8288	0.1955	0.0490
No log	70.0	490	0.5012	0.84	0.2631	0.9562	0.8400	0.8322	0.1945	0.0488
No log	71.0	497	0.5018	0.835	0.2635	0.9568	0.835	0.8288	0.1958	0.0491
0.1944	72.0	504	0.5016	0.84	0.2633	0.9562	0.8400	0.8322	0.2034	0.0490
0.1944	73.0	511	0.5013	0.84	0.2632	0.9558	0.8400	0.8322	0.1944	0.0488
0.1944	74.0	518	0.5013	0.84	0.2631	0.9562	0.8400	0.8322	0.1943	0.0487
0.1944	75.0	525	0.5016	0.835	0.2633	0.9560	0.835	0.8344	0.2035	0.0495
0.1944	76.0	532	0.5018	0.84	0.2634	0.9563	0.8400	0.8322	0.2093	0.0487
0.1944	77.0	539	0.5012	0.84	0.2630	0.9565	0.8400	0.8322	0.1941	0.0488
0.1944	78.0	546	0.5015	0.84	0.2632	0.9561	0.8400	0.8375	0.2008	0.0489
0.1944	79.0	553	0.5016	0.835	0.2633	0.9560	0.835	0.8288	0.1957	0.0490
0.1944	80.0	560	0.5015	0.84	0.2631	0.9568	0.8400	0.8375	0.2093	0.0488
0.1944	81.0	567	0.5015	0.835	0.2632	0.9561	0.835	0.8288	0.1957	0.0491
0.1944	82.0	574	0.5014	0.835	0.2631	0.9565	0.835	0.8288	0.1949	0.0489
0.1944	83.0	581	0.5015	0.835	0.2632	0.9563	0.835	0.8288	0.1957	0.0490
0.1944	84.0	588	0.5015	0.84	0.2632	0.9559	0.8400	0.8322	0.2031	0.0488
0.1944	85.0	595	0.5012	0.84	0.2630	0.9560	0.8400	0.8322	0.1944	0.0488
0.1944	86.0	602	0.5012	0.84	0.2630	0.9561	0.8400	0.8322	0.1944	0.0488
0.1944	87.0	609	0.5012	0.84	0.2630	0.9562	0.8400	0.8322	0.1943	0.0488
0.1944	88.0	616	0.5012	0.84	0.2630	0.9561	0.8400	0.8322	0.2030	0.0488
0.1944	89.0	623	0.5013	0.845	0.2631	0.9559	0.845	0.8409	0.1995	0.0488
0.1944	90.0	630	0.5013	0.845	0.2631	0.9559	0.845	0.8409	0.1995	0.0488
0.1944	91.0	637	0.5012	0.845	0.2630	0.9559	0.845	0.8409	0.1994	0.0487
0.1944	92.0	644	0.5013	0.845	0.2631	0.9561	0.845	0.8409	0.1995	0.0487
0.1944	93.0	651	0.5012	0.84	0.2630	0.9560	0.8400	0.8322	0.2031	0.0488
0.1944	94.0	658	0.5013	0.84	0.2630	0.9558	0.8400	0.8322	0.1944	0.0488
0.1944	95.0	665	0.5012	0.84	0.2630	0.9558	0.8400	0.8322	0.1944	0.0488
0.1944	96.0	672	0.5012	0.84	0.2630	0.9558	0.8400	0.8322	0.1944	0.0488
0.1944	97.0	679	0.5012	0.845	0.2630	0.9559	0.845	0.8409	0.1994	0.0487
0.1944	98.0	686	0.5012	0.84	0.2630	0.9559	0.8400	0.8322	0.1944	0.0488
0.1944	99.0	693	0.5012	0.845	0.2630	0.9560	0.845	0.8409	0.2081	0.0487
0.1944	100.0	700	0.5012	0.845	0.2630	0.9559	0.845	0.8409	0.2081	0.0487

Framework versions

Transformers 4.26.1
Pytorch 1.13.1.post200
Datasets 2.9.0
Tokenizers 0.13.2

jordyvl/vit-small_tobacco3482_kd_CEKD_t2.5_a0.7

vit-small_tobacco3482_kd_CEKD_t2.5_a0.7

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions