dit-base_tobacco-tiny_tobacco3482_kd

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: 0.4787
Accuracy: 0.815
Brier Loss: 0.2625
Nll: 1.3204
F1 Micro: 0.815
F1 Macro: 0.8058
Ece: 0.1408
Aurc: 0.0457

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	2.2200	0.225	0.9173	7.9176	0.225	0.1043	0.3247	0.7528
No log	2.0	14	1.7791	0.37	0.7982	4.5330	0.37	0.2759	0.2753	0.6029
No log	3.0	21	1.3855	0.485	0.6628	2.8048	0.485	0.4006	0.2716	0.3216
No log	4.0	28	1.0594	0.575	0.5416	1.7471	0.575	0.5296	0.2314	0.2052
No log	5.0	35	0.8946	0.645	0.4739	1.6979	0.645	0.6319	0.2168	0.1476
No log	6.0	42	0.8969	0.66	0.4688	1.7181	0.66	0.6621	0.1985	0.1439
No log	7.0	49	0.7466	0.72	0.3937	1.5484	0.72	0.7222	0.2135	0.1079
No log	8.0	56	0.7075	0.725	0.3753	1.4585	0.7250	0.7100	0.2029	0.0894
No log	9.0	63	0.9838	0.695	0.4504	2.0243	0.695	0.6681	0.2306	0.1204
No log	10.0	70	0.7683	0.73	0.3867	1.5378	0.7300	0.7388	0.1699	0.0902
No log	11.0	77	0.7114	0.755	0.3524	1.5684	0.755	0.7296	0.1687	0.0742
No log	12.0	84	0.8151	0.76	0.3728	1.4721	0.76	0.7336	0.1863	0.0936
No log	13.0	91	0.8346	0.77	0.3524	1.9540	0.7700	0.7528	0.1691	0.0790
No log	14.0	98	0.7822	0.735	0.3898	1.6873	0.735	0.7215	0.2025	0.0860
No log	15.0	105	0.7400	0.765	0.3580	1.5692	0.765	0.7167	0.1765	0.0809
No log	16.0	112	0.8296	0.71	0.4027	1.5508	0.7100	0.6837	0.2185	0.0963
No log	17.0	119	0.6868	0.79	0.3443	1.4135	0.79	0.7528	0.1971	0.0748
No log	18.0	126	0.6290	0.795	0.3142	1.8030	0.795	0.7657	0.1476	0.0621
No log	19.0	133	0.6454	0.79	0.3213	1.6664	0.79	0.7785	0.1649	0.0645
No log	20.0	140	0.6089	0.785	0.3228	1.5436	0.785	0.7729	0.1708	0.0639
No log	21.0	147	0.6715	0.785	0.3289	1.3422	0.785	0.7598	0.1787	0.0768
No log	22.0	154	0.7075	0.79	0.3342	1.6069	0.79	0.7684	0.1587	0.0656
No log	23.0	161	0.6226	0.805	0.3067	1.2400	0.805	0.7881	0.1611	0.0716
No log	24.0	168	0.7501	0.77	0.3506	1.8952	0.7700	0.7530	0.1637	0.0746
No log	25.0	175	0.6039	0.775	0.3168	1.4196	0.775	0.7647	0.1701	0.0664
No log	26.0	182	0.6252	0.775	0.3260	1.4914	0.775	0.7507	0.1733	0.0657
No log	27.0	189	0.6590	0.79	0.3303	1.5970	0.79	0.7773	0.1695	0.0747
No log	28.0	196	0.5920	0.815	0.2988	1.6841	0.815	0.8127	0.1711	0.0635
No log	29.0	203	0.5982	0.785	0.3163	1.6290	0.785	0.7678	0.1641	0.0597
No log	30.0	210	0.5693	0.805	0.3028	1.4954	0.805	0.7917	0.1566	0.0578
No log	31.0	217	0.5860	0.805	0.2964	1.3856	0.805	0.7966	0.1413	0.0599
No log	32.0	224	0.5380	0.805	0.2775	1.6946	0.805	0.7981	0.1526	0.0494
No log	33.0	231	0.5041	0.8	0.2745	1.6025	0.8000	0.7887	0.1639	0.0498
No log	34.0	238	0.5134	0.83	0.2700	1.3768	0.83	0.8161	0.1464	0.0526
No log	35.0	245	0.5371	0.81	0.2820	1.3584	0.81	0.7982	0.1466	0.0552
No log	36.0	252	0.4987	0.815	0.2711	1.3735	0.815	0.8056	0.1540	0.0490
No log	37.0	259	0.5145	0.81	0.2814	1.3537	0.81	0.8000	0.1415	0.0521
No log	38.0	266	0.4992	0.815	0.2721	1.3420	0.815	0.7974	0.1453	0.0497
No log	39.0	273	0.4992	0.795	0.2748	1.3579	0.795	0.7757	0.1485	0.0502
No log	40.0	280	0.4881	0.82	0.2634	1.3745	0.82	0.8058	0.1504	0.0475
No log	41.0	287	0.4977	0.81	0.2750	1.3208	0.81	0.7965	0.1520	0.0504
No log	42.0	294	0.4865	0.815	0.2644	1.3840	0.815	0.8056	0.1517	0.0452
No log	43.0	301	0.5034	0.81	0.2722	1.3683	0.81	0.7967	0.1404	0.0514
No log	44.0	308	0.4925	0.815	0.2692	1.3979	0.815	0.8056	0.1386	0.0462
No log	45.0	315	0.4643	0.825	0.2608	1.3015	0.825	0.8148	0.1516	0.0442
No log	46.0	322	0.4851	0.82	0.2666	1.2561	0.82	0.8018	0.1494	0.0461
No log	47.0	329	0.4751	0.82	0.2615	1.3167	0.82	0.8120	0.1544	0.0457
No log	48.0	336	0.4666	0.82	0.2596	1.2470	0.82	0.8120	0.1326	0.0443
No log	49.0	343	0.4856	0.815	0.2659	1.3283	0.815	0.8081	0.1501	0.0474
No log	50.0	350	0.4690	0.83	0.2618	1.3227	0.83	0.8208	0.1448	0.0435
No log	51.0	357	0.4835	0.81	0.2670	1.2956	0.81	0.7961	0.1425	0.0471
No log	52.0	364	0.4857	0.82	0.2598	1.3115	0.82	0.8134	0.1547	0.0478
No log	53.0	371	0.4747	0.82	0.2654	1.3717	0.82	0.8026	0.1596	0.0453
No log	54.0	378	0.4925	0.815	0.2649	1.2289	0.815	0.8058	0.1452	0.0486
No log	55.0	385	0.4670	0.825	0.2611	1.3080	0.825	0.8102	0.1307	0.0434
No log	56.0	392	0.4878	0.81	0.2636	1.3040	0.81	0.7961	0.1546	0.0478
No log	57.0	399	0.4679	0.82	0.2600	1.2618	0.82	0.8038	0.1516	0.0430
No log	58.0	406	0.4802	0.815	0.2629	1.3054	0.815	0.8079	0.1448	0.0476
No log	59.0	413	0.4746	0.82	0.2615	1.3177	0.82	0.8064	0.1308	0.0457
No log	60.0	420	0.4784	0.82	0.2608	1.2495	0.82	0.8134	0.1336	0.0463
No log	61.0	427	0.4751	0.82	0.2630	1.2886	0.82	0.8086	0.1416	0.0459
No log	62.0	434	0.4751	0.815	0.2606	1.2453	0.815	0.8058	0.1529	0.0455
No log	63.0	441	0.4737	0.825	0.2629	1.2975	0.825	0.8113	0.1286	0.0451
No log	64.0	448	0.4840	0.815	0.2631	1.3210	0.815	0.8036	0.1392	0.0472
No log	65.0	455	0.4747	0.82	0.2615	1.3054	0.82	0.8086	0.1491	0.0456
No log	66.0	462	0.4767	0.815	0.2618	1.3056	0.815	0.8058	0.1517	0.0459
No log	67.0	469	0.4748	0.82	0.2615	1.3046	0.82	0.8086	0.1525	0.0453
No log	68.0	476	0.4782	0.815	0.2626	1.3088	0.815	0.8058	0.1519	0.0461
No log	69.0	483	0.4769	0.815	0.2616	1.3133	0.815	0.8058	0.1555	0.0456
No log	70.0	490	0.4767	0.815	0.2622	1.3067	0.815	0.8058	0.1435	0.0457
No log	71.0	497	0.4776	0.815	0.2623	1.3111	0.815	0.8058	0.1533	0.0458
0.1688	72.0	504	0.4770	0.815	0.2621	1.3078	0.815	0.8058	0.1605	0.0457
0.1688	73.0	511	0.4783	0.815	0.2625	1.3109	0.815	0.8058	0.1503	0.0458
0.1688	74.0	518	0.4776	0.815	0.2621	1.3117	0.815	0.8058	0.1648	0.0458
0.1688	75.0	525	0.4784	0.815	0.2627	1.3110	0.815	0.8058	0.1463	0.0458
0.1688	76.0	532	0.4779	0.815	0.2625	1.3125	0.815	0.8058	0.1577	0.0457
0.1688	77.0	539	0.4794	0.815	0.2628	1.3110	0.815	0.8058	0.1420	0.0459
0.1688	78.0	546	0.4776	0.815	0.2623	1.3120	0.815	0.8058	0.1517	0.0455
0.1688	79.0	553	0.4789	0.815	0.2627	1.3101	0.815	0.8058	0.1460	0.0459
0.1688	80.0	560	0.4784	0.815	0.2626	1.3127	0.815	0.8058	0.1518	0.0457
0.1688	81.0	567	0.4782	0.815	0.2625	1.3103	0.815	0.8058	0.1408	0.0457
0.1688	82.0	574	0.4791	0.815	0.2627	1.3166	0.815	0.8058	0.1586	0.0458
0.1688	83.0	581	0.4785	0.815	0.2625	1.3116	0.815	0.8058	0.1436	0.0459
0.1688	84.0	588	0.4783	0.815	0.2624	1.3113	0.815	0.8058	0.1476	0.0458
0.1688	85.0	595	0.4785	0.815	0.2625	1.3169	0.815	0.8058	0.1500	0.0457
0.1688	86.0	602	0.4782	0.815	0.2625	1.3127	0.815	0.8058	0.1496	0.0457
0.1688	87.0	609	0.4778	0.815	0.2623	1.3119	0.815	0.8058	0.1496	0.0456
0.1688	88.0	616	0.4783	0.815	0.2625	1.3118	0.815	0.8058	0.1529	0.0458
0.1688	89.0	623	0.4784	0.815	0.2625	1.3149	0.815	0.8058	0.1485	0.0457
0.1688	90.0	630	0.4781	0.815	0.2624	1.3137	0.815	0.8058	0.1472	0.0457
0.1688	91.0	637	0.4784	0.815	0.2626	1.3111	0.815	0.8058	0.1492	0.0458
0.1688	92.0	644	0.4785	0.815	0.2625	1.3177	0.815	0.8058	0.1485	0.0457
0.1688	93.0	651	0.4790	0.815	0.2626	1.3208	0.815	0.8058	0.1462	0.0457
0.1688	94.0	658	0.4788	0.815	0.2625	1.3178	0.815	0.8058	0.1396	0.0458
0.1688	95.0	665	0.4785	0.815	0.2625	1.3203	0.815	0.8058	0.1484	0.0457
0.1688	96.0	672	0.4786	0.815	0.2625	1.3168	0.815	0.8058	0.1470	0.0457
0.1688	97.0	679	0.4786	0.815	0.2625	1.3167	0.815	0.8058	0.1470	0.0457
0.1688	98.0	686	0.4787	0.815	0.2625	1.3192	0.815	0.8058	0.1408	0.0457
0.1688	99.0	693	0.4787	0.815	0.2625	1.3205	0.815	0.8058	0.1408	0.0457
0.1688	100.0	700	0.4787	0.815	0.2625	1.3204	0.815	0.8058	0.1408	0.0457

Framework versions

Transformers 4.36.0.dev0
Pytorch 2.2.0.dev20231112+cu118
Datasets 2.14.5
Tokenizers 0.14.1

jordyvl/dit-base_tobacco-tiny_tobacco3482_kd

dit-base_tobacco-tiny_tobacco3482_kd

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions