81-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: 0.4383
Accuracy: 0.815
Brier Loss: 0.3137
Nll: 0.9180
F1 Micro: 0.815
F1 Macro: 0.7935
Ece: 0.2739
Aurc: 0.0546

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	2.1414	0.1	1.0055	8.0967	0.1000	0.0761	0.3253	0.8939
No log	2.0	14	1.4539	0.215	0.8614	5.0189	0.2150	0.1812	0.2795	0.7550
No log	3.0	21	1.1486	0.39	0.7467	4.3499	0.39	0.3077	0.2755	0.4066
No log	4.0	28	0.9182	0.575	0.5990	2.9424	0.575	0.4619	0.2988	0.2367
No log	5.0	35	0.8134	0.645	0.5398	2.1261	0.645	0.5549	0.2689	0.1878
No log	6.0	42	0.7149	0.7	0.4612	2.2189	0.7	0.6006	0.2518	0.1316
No log	7.0	49	0.6607	0.73	0.4303	1.7076	0.7300	0.6689	0.2976	0.1200
No log	8.0	56	0.6109	0.755	0.3922	1.7703	0.755	0.6743	0.2638	0.0861
No log	9.0	63	0.6119	0.715	0.3943	1.4194	0.715	0.6534	0.2276	0.0840
No log	10.0	70	0.5462	0.8	0.3535	1.4026	0.8000	0.7760	0.2888	0.0660
No log	11.0	77	0.5376	0.785	0.3481	1.2329	0.785	0.7600	0.2660	0.0757
No log	12.0	84	0.5250	0.785	0.3442	1.1226	0.785	0.7669	0.2476	0.0802
No log	13.0	91	0.5053	0.81	0.3313	1.1038	0.81	0.7824	0.2543	0.0628
No log	14.0	98	0.5188	0.79	0.3497	1.0872	0.79	0.7678	0.2495	0.0759
No log	15.0	105	0.5020	0.805	0.3412	1.4342	0.805	0.7868	0.2669	0.0652
No log	16.0	112	0.5221	0.795	0.3496	1.3609	0.795	0.7473	0.2682	0.0621
No log	17.0	119	0.5046	0.8	0.3372	1.1543	0.8000	0.7766	0.2604	0.0689
No log	18.0	126	0.4733	0.805	0.3248	1.1335	0.805	0.7800	0.2599	0.0582
No log	19.0	133	0.4725	0.815	0.3242	1.1607	0.815	0.7855	0.2733	0.0573
No log	20.0	140	0.4887	0.82	0.3325	1.0316	0.82	0.7998	0.2684	0.0568
No log	21.0	147	0.4708	0.815	0.3205	1.1275	0.815	0.8033	0.2412	0.0663
No log	22.0	154	0.4773	0.83	0.3309	1.1147	0.83	0.8101	0.2739	0.0518
No log	23.0	161	0.4957	0.815	0.3402	1.0884	0.815	0.8012	0.2582	0.0726
No log	24.0	168	0.4666	0.805	0.3305	1.0784	0.805	0.7858	0.2792	0.0560
No log	25.0	175	0.4830	0.795	0.3324	1.1757	0.795	0.7595	0.2505	0.0715
No log	26.0	182	0.4622	0.8	0.3272	1.0698	0.8000	0.7873	0.2795	0.0590
No log	27.0	189	0.4604	0.8	0.3200	1.1104	0.8000	0.7717	0.2561	0.0630
No log	28.0	196	0.4635	0.82	0.3253	1.1271	0.82	0.7903	0.2756	0.0571
No log	29.0	203	0.4590	0.815	0.3211	1.1048	0.815	0.7952	0.2881	0.0567
No log	30.0	210	0.4575	0.795	0.3210	0.9174	0.795	0.7786	0.2833	0.0625
No log	31.0	217	0.4684	0.83	0.3337	0.9485	0.83	0.8093	0.2892	0.0557
No log	32.0	224	0.4520	0.81	0.3208	1.0186	0.81	0.7955	0.2438	0.0577
No log	33.0	231	0.4567	0.825	0.3233	0.9246	0.825	0.7928	0.2665	0.0592
No log	34.0	238	0.4468	0.82	0.3152	1.0065	0.82	0.8000	0.2710	0.0563
No log	35.0	245	0.4562	0.78	0.3244	1.0626	0.78	0.7614	0.2602	0.0624
No log	36.0	252	0.4542	0.815	0.3223	1.1362	0.815	0.7852	0.2584	0.0579
No log	37.0	259	0.4441	0.82	0.3136	1.0419	0.82	0.7901	0.2790	0.0529
No log	38.0	266	0.4408	0.825	0.3125	0.9860	0.825	0.8023	0.2766	0.0553
No log	39.0	273	0.4354	0.83	0.3082	0.8958	0.83	0.8116	0.2713	0.0504
No log	40.0	280	0.4465	0.79	0.3164	1.1111	0.79	0.7715	0.2668	0.0628
No log	41.0	287	0.4416	0.845	0.3128	1.0103	0.845	0.8162	0.3044	0.0527
No log	42.0	294	0.4463	0.83	0.3165	1.0849	0.83	0.8106	0.2683	0.0580
No log	43.0	301	0.4405	0.845	0.3132	1.0312	0.845	0.8247	0.2792	0.0509
No log	44.0	308	0.4443	0.83	0.3174	0.9196	0.83	0.8094	0.2687	0.0524
No log	45.0	315	0.4445	0.82	0.3194	1.0665	0.82	0.7897	0.2651	0.0560
No log	46.0	322	0.4405	0.81	0.3133	1.1805	0.81	0.7770	0.2771	0.0550
No log	47.0	329	0.4380	0.84	0.3132	0.9508	0.8400	0.8104	0.2916	0.0535
No log	48.0	336	0.4407	0.825	0.3139	0.9044	0.825	0.7978	0.2702	0.0542
No log	49.0	343	0.4418	0.835	0.3154	0.8965	0.835	0.8178	0.2877	0.0569
No log	50.0	350	0.4368	0.825	0.3123	0.9774	0.825	0.8073	0.2607	0.0531
No log	51.0	357	0.4402	0.825	0.3140	0.9170	0.825	0.8052	0.2810	0.0550
No log	52.0	364	0.4374	0.82	0.3107	0.9873	0.82	0.7952	0.2602	0.0542
No log	53.0	371	0.4368	0.83	0.3120	0.9832	0.83	0.8084	0.2709	0.0541
No log	54.0	378	0.4375	0.82	0.3131	0.9094	0.82	0.7943	0.2633	0.0538
No log	55.0	385	0.4379	0.815	0.3134	0.9927	0.815	0.7856	0.2960	0.0552
No log	56.0	392	0.4370	0.83	0.3125	0.9746	0.83	0.8100	0.2744	0.0535
No log	57.0	399	0.4366	0.825	0.3123	1.0392	0.825	0.8021	0.2730	0.0536
No log	58.0	406	0.4372	0.825	0.3129	0.9174	0.825	0.8026	0.2800	0.0542
No log	59.0	413	0.4380	0.81	0.3134	0.9770	0.81	0.7831	0.2612	0.0557
No log	60.0	420	0.4374	0.82	0.3130	0.9124	0.82	0.7961	0.2589	0.0541
No log	61.0	427	0.4366	0.825	0.3121	0.9038	0.825	0.8061	0.2641	0.0538
No log	62.0	434	0.4372	0.825	0.3126	0.9105	0.825	0.8042	0.2684	0.0547
No log	63.0	441	0.4381	0.82	0.3135	0.9160	0.82	0.7961	0.2810	0.0545
No log	64.0	448	0.4376	0.83	0.3133	0.9134	0.83	0.8100	0.2757	0.0539
No log	65.0	455	0.4376	0.825	0.3130	0.9133	0.825	0.8061	0.2977	0.0541
No log	66.0	462	0.4378	0.825	0.3133	0.9153	0.825	0.8061	0.2767	0.0543
No log	67.0	469	0.4373	0.825	0.3129	0.9139	0.825	0.8042	0.2905	0.0541
No log	68.0	476	0.4375	0.82	0.3129	0.9128	0.82	0.7961	0.2739	0.0543
No log	69.0	483	0.4376	0.82	0.3131	0.9125	0.82	0.7961	0.2757	0.0542
No log	70.0	490	0.4377	0.825	0.3133	0.9174	0.825	0.8061	0.2924	0.0538
No log	71.0	497	0.4380	0.82	0.3134	0.9179	0.82	0.7961	0.2896	0.0541
0.2684	72.0	504	0.4378	0.82	0.3133	0.9148	0.82	0.8035	0.2912	0.0543
0.2684	73.0	511	0.4375	0.82	0.3131	0.9169	0.82	0.7961	0.2731	0.0542
0.2684	74.0	518	0.4379	0.82	0.3133	0.9177	0.82	0.7961	0.2732	0.0540
0.2684	75.0	525	0.4383	0.82	0.3138	0.9194	0.82	0.8035	0.2835	0.0545
0.2684	76.0	532	0.4378	0.815	0.3133	0.9133	0.815	0.7935	0.2824	0.0543
0.2684	77.0	539	0.4378	0.815	0.3133	0.9146	0.815	0.7935	0.2735	0.0542
0.2684	78.0	546	0.4379	0.815	0.3134	0.9139	0.815	0.7935	0.2828	0.0547
0.2684	79.0	553	0.4382	0.815	0.3136	0.9179	0.815	0.7935	0.2817	0.0547
0.2684	80.0	560	0.4380	0.815	0.3134	0.9168	0.815	0.7935	0.2818	0.0545
0.2684	81.0	567	0.4381	0.815	0.3135	0.9183	0.815	0.7935	0.2736	0.0544
0.2684	82.0	574	0.4379	0.815	0.3134	0.9164	0.815	0.7935	0.2736	0.0544
0.2684	83.0	581	0.4382	0.815	0.3136	0.9168	0.815	0.7935	0.2736	0.0541
0.2684	84.0	588	0.4381	0.815	0.3136	0.9199	0.815	0.7935	0.2737	0.0541
0.2684	85.0	595	0.4380	0.815	0.3134	0.9175	0.815	0.7935	0.2735	0.0543
0.2684	86.0	602	0.4383	0.815	0.3137	0.9197	0.815	0.7935	0.2674	0.0545
0.2684	87.0	609	0.4381	0.815	0.3135	0.9176	0.815	0.7935	0.2738	0.0547
0.2684	88.0	616	0.4381	0.815	0.3135	0.9179	0.815	0.7935	0.2736	0.0541
0.2684	89.0	623	0.4381	0.815	0.3135	0.9177	0.815	0.7935	0.2737	0.0543
0.2684	90.0	630	0.4382	0.815	0.3136	0.9173	0.815	0.7935	0.2736	0.0544
0.2684	91.0	637	0.4383	0.815	0.3136	0.9187	0.815	0.7935	0.2738	0.0546
0.2684	92.0	644	0.4382	0.815	0.3136	0.9171	0.815	0.7935	0.2738	0.0544
0.2684	93.0	651	0.4383	0.815	0.3137	0.9190	0.815	0.7935	0.2738	0.0546
0.2684	94.0	658	0.4382	0.815	0.3136	0.9187	0.815	0.7935	0.2737	0.0543
0.2684	95.0	665	0.4383	0.815	0.3136	0.9184	0.815	0.7935	0.2739	0.0545
0.2684	96.0	672	0.4382	0.815	0.3136	0.9184	0.815	0.7935	0.2737	0.0545
0.2684	97.0	679	0.4382	0.815	0.3136	0.9179	0.815	0.7935	0.2739	0.0545
0.2684	98.0	686	0.4383	0.815	0.3137	0.9185	0.815	0.7935	0.2738	0.0544
0.2684	99.0	693	0.4383	0.815	0.3137	0.9182	0.815	0.7935	0.2739	0.0546
0.2684	100.0	700	0.4383	0.815	0.3137	0.9180	0.815	0.7935	0.2739	0.0546

Framework versions

Transformers 4.26.1
Pytorch 1.13.1.post200
Datasets 2.9.0
Tokenizers 0.13.2

81-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