vit-tiny_tobacco3482_kd_MSE

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: 0.2482
Accuracy: 0.82
Brier Loss: 0.3226
Nll: 0.8343
F1 Micro: 0.82
F1 Macro: 0.8090
Ece: 0.2625
Aurc: 0.0606

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	2.6314	0.165	1.0293	8.4704	0.165	0.0552	0.3856	0.8464
No log	2.0	14	1.4800	0.12	0.9052	7.3409	0.12	0.0997	0.2469	0.8374
No log	3.0	21	1.0627	0.34	0.8400	5.5369	0.34	0.2006	0.3075	0.5758
No log	4.0	28	0.8429	0.42	0.7462	3.2442	0.4200	0.3362	0.3072	0.3692
No log	5.0	35	0.7060	0.535	0.6558	2.7042	0.535	0.4295	0.2993	0.2677
No log	6.0	42	0.5950	0.635	0.6056	2.0779	0.635	0.5021	0.3375	0.1948
No log	7.0	49	0.4865	0.67	0.5486	1.4919	0.67	0.5384	0.3633	0.1737
No log	8.0	56	0.4572	0.69	0.4897	1.4359	0.69	0.6106	0.2889	0.1355
No log	9.0	63	0.3932	0.72	0.4496	1.0748	0.72	0.6261	0.2914	0.1092
No log	10.0	70	0.3584	0.76	0.4091	1.1341	0.76	0.6747	0.2946	0.0937
No log	11.0	77	0.3516	0.785	0.3906	1.0586	0.785	0.7422	0.3026	0.0762
No log	12.0	84	0.3905	0.74	0.4155	1.1502	0.74	0.6677	0.2827	0.1211
No log	13.0	91	0.3346	0.775	0.3640	1.0888	0.775	0.7397	0.2743	0.0771
No log	14.0	98	0.3700	0.81	0.3728	0.9575	0.81	0.7941	0.3125	0.0709
No log	15.0	105	0.3346	0.8	0.3631	0.9636	0.8000	0.7731	0.3113	0.0662
No log	16.0	112	0.3084	0.785	0.3606	1.0243	0.785	0.7610	0.2883	0.0911
No log	17.0	119	0.3266	0.785	0.3375	1.0301	0.785	0.7603	0.2659	0.0687
No log	18.0	126	0.2987	0.805	0.3371	0.7961	0.805	0.7895	0.2764	0.0639
No log	19.0	133	0.3468	0.815	0.3428	0.9841	0.815	0.7884	0.2824	0.0641
No log	20.0	140	0.3111	0.81	0.3409	0.8873	0.81	0.8010	0.2978	0.0579
No log	21.0	147	0.3042	0.8	0.3392	0.8057	0.8000	0.7802	0.2611	0.0556
No log	22.0	154	0.2936	0.84	0.3359	0.8659	0.8400	0.8210	0.2920	0.0636
No log	23.0	161	0.3284	0.815	0.3457	0.9494	0.815	0.8007	0.2793	0.0615
No log	24.0	168	0.3184	0.805	0.3474	0.9087	0.805	0.7851	0.2793	0.0605
No log	25.0	175	0.3081	0.805	0.3357	1.0425	0.805	0.7757	0.2781	0.0646
No log	26.0	182	0.2743	0.815	0.3425	0.7898	0.815	0.8028	0.2814	0.0689
No log	27.0	189	0.2873	0.81	0.3254	0.8795	0.81	0.7904	0.2755	0.0637
No log	28.0	196	0.2650	0.82	0.3236	0.7462	0.82	0.7986	0.2714	0.0616
No log	29.0	203	0.2756	0.815	0.3310	0.8227	0.815	0.7973	0.2766	0.0664
No log	30.0	210	0.2711	0.83	0.3337	0.8340	0.83	0.8186	0.2992	0.0688
No log	31.0	217	0.2779	0.825	0.3234	0.8173	0.825	0.8100	0.2579	0.0589
No log	32.0	224	0.2679	0.82	0.3216	0.7441	0.82	0.8024	0.2940	0.0605
No log	33.0	231	0.2633	0.805	0.3277	0.8046	0.805	0.7871	0.2710	0.0644
No log	34.0	238	0.2705	0.805	0.3333	0.8661	0.805	0.7890	0.2626	0.0632
No log	35.0	245	0.2624	0.815	0.3295	0.8568	0.815	0.7900	0.2865	0.0651
No log	36.0	252	0.2654	0.805	0.3262	0.8075	0.805	0.7793	0.2726	0.0662
No log	37.0	259	0.2697	0.805	0.3293	0.8143	0.805	0.7857	0.2587	0.0623
No log	38.0	266	0.2548	0.805	0.3267	0.8028	0.805	0.7847	0.2606	0.0660
No log	39.0	273	0.2740	0.83	0.3218	0.8270	0.83	0.8172	0.2697	0.0566
No log	40.0	280	0.2572	0.81	0.3302	0.8573	0.81	0.7892	0.2722	0.0663
No log	41.0	287	0.2528	0.81	0.3300	0.8454	0.81	0.7980	0.2555	0.0673
No log	42.0	294	0.2590	0.815	0.3271	0.8393	0.815	0.8002	0.2554	0.0604
No log	43.0	301	0.2654	0.825	0.3273	0.8100	0.825	0.8155	0.2687	0.0595
No log	44.0	308	0.2506	0.805	0.3290	0.8551	0.805	0.7858	0.2425	0.0716
No log	45.0	315	0.2615	0.82	0.3255	0.8624	0.82	0.8007	0.2773	0.0586
No log	46.0	322	0.2487	0.815	0.3240	0.8416	0.815	0.7987	0.2757	0.0611
No log	47.0	329	0.2674	0.845	0.3207	0.8407	0.845	0.8227	0.2963	0.0553
No log	48.0	336	0.2522	0.805	0.3312	0.9376	0.805	0.7888	0.2714	0.0662
No log	49.0	343	0.2547	0.81	0.3280	0.7847	0.81	0.7870	0.2696	0.0688
No log	50.0	350	0.2523	0.81	0.3213	0.7968	0.81	0.7937	0.2599	0.0654
No log	51.0	357	0.2526	0.815	0.3291	0.8022	0.815	0.7994	0.2888	0.0669
No log	52.0	364	0.2568	0.835	0.3180	0.8317	0.835	0.8140	0.2750	0.0578
No log	53.0	371	0.2496	0.82	0.3267	0.8442	0.82	0.8068	0.2825	0.0602
No log	54.0	378	0.2602	0.82	0.3229	0.7963	0.82	0.8061	0.2714	0.0585
No log	55.0	385	0.2477	0.81	0.3237	0.8278	0.81	0.7937	0.2511	0.0631
No log	56.0	392	0.2508	0.83	0.3210	0.8302	0.83	0.8102	0.2706	0.0588
No log	57.0	399	0.2454	0.815	0.3240	0.8377	0.815	0.7948	0.2607	0.0643
No log	58.0	406	0.2488	0.815	0.3229	0.8308	0.815	0.7942	0.2569	0.0604
No log	59.0	413	0.2510	0.82	0.3223	0.8314	0.82	0.8093	0.2779	0.0603
No log	60.0	420	0.2499	0.82	0.3235	0.8401	0.82	0.8031	0.2578	0.0618
No log	61.0	427	0.2478	0.81	0.3227	0.8315	0.81	0.7933	0.2645	0.0615
No log	62.0	434	0.2460	0.82	0.3231	0.8364	0.82	0.8028	0.2795	0.0643
No log	63.0	441	0.2489	0.825	0.3224	0.8337	0.825	0.8156	0.2759	0.0604
No log	64.0	448	0.2482	0.825	0.3230	0.8320	0.825	0.8138	0.2753	0.0600
No log	65.0	455	0.2462	0.815	0.3231	0.8354	0.815	0.8008	0.2551	0.0625
No log	66.0	462	0.2470	0.815	0.3219	0.8338	0.815	0.8018	0.2729	0.0611
No log	67.0	469	0.2457	0.81	0.3231	0.8336	0.81	0.7930	0.2587	0.0638
No log	68.0	476	0.2472	0.815	0.3225	0.8334	0.815	0.8008	0.2706	0.0619
No log	69.0	483	0.2473	0.825	0.3223	0.8357	0.825	0.8165	0.2668	0.0611
No log	70.0	490	0.2481	0.81	0.3223	0.8343	0.81	0.7930	0.2685	0.0624
No log	71.0	497	0.2472	0.825	0.3227	0.8338	0.825	0.8117	0.2839	0.0601
0.232	72.0	504	0.2472	0.815	0.3220	0.8345	0.815	0.8018	0.2617	0.0615
0.232	73.0	511	0.2486	0.82	0.3218	0.8321	0.82	0.8086	0.2768	0.0610
0.232	74.0	518	0.2468	0.815	0.3219	0.8338	0.815	0.8008	0.2717	0.0621
0.232	75.0	525	0.2470	0.82	0.3223	0.8325	0.82	0.8090	0.2625	0.0606
0.232	76.0	532	0.2474	0.825	0.3223	0.8322	0.825	0.8165	0.2723	0.0602
0.232	77.0	539	0.2476	0.805	0.3227	0.8345	0.805	0.7859	0.2589	0.0629
0.232	78.0	546	0.2479	0.82	0.3228	0.8336	0.82	0.8090	0.2674	0.0608
0.232	79.0	553	0.2478	0.82	0.3225	0.8349	0.82	0.8090	0.2624	0.0604
0.232	80.0	560	0.2477	0.81	0.3227	0.8337	0.81	0.7938	0.2577	0.0621
0.232	81.0	567	0.2478	0.82	0.3226	0.8336	0.82	0.8090	0.2670	0.0607
0.232	82.0	574	0.2480	0.825	0.3224	0.8340	0.825	0.8165	0.2673	0.0605
0.232	83.0	581	0.2479	0.82	0.3227	0.8347	0.82	0.8090	0.2564	0.0607
0.232	84.0	588	0.2480	0.82	0.3226	0.8342	0.82	0.8090	0.2625	0.0606
0.232	85.0	595	0.2480	0.82	0.3225	0.8339	0.82	0.8090	0.2625	0.0606
0.232	86.0	602	0.2479	0.825	0.3226	0.8339	0.825	0.8165	0.2677	0.0606
0.232	87.0	609	0.2479	0.82	0.3225	0.8339	0.82	0.8090	0.2624	0.0605
0.232	88.0	616	0.2481	0.825	0.3225	0.8343	0.825	0.8165	0.2675	0.0604
0.232	89.0	623	0.2481	0.825	0.3225	0.8341	0.825	0.8165	0.2722	0.0605
0.232	90.0	630	0.2481	0.82	0.3225	0.8341	0.82	0.8090	0.2625	0.0606
0.232	91.0	637	0.2481	0.82	0.3226	0.8345	0.82	0.8090	0.2629	0.0608
0.232	92.0	644	0.2481	0.825	0.3226	0.8342	0.825	0.8165	0.2675	0.0605
0.232	93.0	651	0.2481	0.825	0.3225	0.8340	0.825	0.8165	0.2675	0.0605
0.232	94.0	658	0.2481	0.82	0.3225	0.8343	0.82	0.8090	0.2625	0.0606
0.232	95.0	665	0.2482	0.82	0.3226	0.8345	0.82	0.8090	0.2627	0.0606
0.232	96.0	672	0.2482	0.82	0.3225	0.8343	0.82	0.8090	0.2627	0.0607
0.232	97.0	679	0.2482	0.82	0.3226	0.8344	0.82	0.8090	0.2627	0.0607
0.232	98.0	686	0.2482	0.82	0.3226	0.8344	0.82	0.8090	0.2626	0.0606
0.232	99.0	693	0.2482	0.82	0.3226	0.8343	0.82	0.8090	0.2625	0.0606
0.232	100.0	700	0.2482	0.82	0.3226	0.8343	0.82	0.8090	0.2625	0.0606

Framework versions

Transformers 4.26.1
Pytorch 1.13.1.post200
Datasets 2.9.0
Tokenizers 0.13.2

jordyvl/vit-tiny_tobacco3482_kd_MSE

vit-tiny_tobacco3482_kd_MSE

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions