vit-tiny_tobacco3482_kd_NKD_t1.0_g1.5

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: 3.6948
Accuracy: 0.85
Brier Loss: 0.2427
Nll: 1.2265
F1 Micro: 0.85
F1 Macro: 0.8401
Ece: 0.1325
Aurc: 0.0510

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	6.1834	0.09	1.0218	8.8168	0.09	0.0691	0.3227	0.8979
No log	2.0	14	5.0430	0.14	0.9049	8.5503	0.14	0.1102	0.2495	0.7963
No log	3.0	21	4.6038	0.26	0.8385	6.2625	0.26	0.1872	0.2701	0.5801
No log	4.0	28	4.3342	0.475	0.7283	4.0652	0.4750	0.3906	0.3309	0.3230
No log	5.0	35	4.0844	0.575	0.6168	2.7058	0.575	0.4691	0.2963	0.2299
No log	6.0	42	3.9551	0.67	0.5721	2.2402	0.67	0.5531	0.3412	0.1586
No log	7.0	49	3.8114	0.695	0.4999	1.8621	0.695	0.5789	0.3028	0.1342
No log	8.0	56	3.7627	0.77	0.4472	1.6313	0.7700	0.7306	0.3170	0.0894
No log	9.0	63	3.7510	0.75	0.4223	1.4350	0.75	0.6574	0.2576	0.1046
No log	10.0	70	3.6586	0.8	0.3576	1.5473	0.8000	0.7615	0.2754	0.0701
No log	11.0	77	3.6643	0.825	0.3327	1.4335	0.825	0.7921	0.2555	0.0670
No log	12.0	84	3.6561	0.79	0.3465	1.3859	0.79	0.7541	0.2165	0.0833
No log	13.0	91	3.6518	0.805	0.3209	1.0963	0.805	0.7716	0.1894	0.0706
No log	14.0	98	3.6401	0.84	0.3045	1.1974	0.8400	0.8290	0.2261	0.0551
No log	15.0	105	3.6253	0.79	0.3200	1.2187	0.79	0.7658	0.1973	0.0777
No log	16.0	112	3.6019	0.83	0.2920	1.0547	0.83	0.8050	0.1655	0.0632
No log	17.0	119	3.5622	0.84	0.2741	1.0769	0.8400	0.8228	0.1641	0.0636
No log	18.0	126	3.5921	0.835	0.2706	1.0996	0.835	0.8205	0.1810	0.0581
No log	19.0	133	3.5886	0.815	0.2773	1.2384	0.815	0.7975	0.1772	0.0643
No log	20.0	140	3.5798	0.84	0.2732	1.4226	0.8400	0.8236	0.1624	0.0672
No log	21.0	147	3.5683	0.82	0.2868	1.3978	0.82	0.7990	0.1629	0.0681
No log	22.0	154	3.5891	0.825	0.2893	1.4084	0.825	0.8024	0.1856	0.0666
No log	23.0	161	3.5484	0.82	0.2595	1.0782	0.82	0.7990	0.1477	0.0595
No log	24.0	168	3.5882	0.8	0.2686	1.0513	0.8000	0.7745	0.1334	0.0596
No log	25.0	175	3.5636	0.81	0.2774	1.1250	0.81	0.7838	0.1563	0.0662
No log	26.0	182	3.5478	0.82	0.2669	1.0724	0.82	0.8021	0.1318	0.0576
No log	27.0	189	3.5092	0.84	0.2487	1.0652	0.8400	0.8144	0.1500	0.0567
No log	28.0	196	3.5746	0.815	0.2810	1.3942	0.815	0.7997	0.1547	0.0682
No log	29.0	203	3.5633	0.835	0.2587	1.1580	0.835	0.8154	0.1476	0.0604
No log	30.0	210	3.5221	0.835	0.2591	0.9385	0.835	0.8180	0.1496	0.0584
No log	31.0	217	3.6263	0.83	0.2676	1.3260	0.83	0.8213	0.1670	0.0550
No log	32.0	224	3.5758	0.825	0.2855	1.3100	0.825	0.8082	0.1453	0.0637
No log	33.0	231	3.5836	0.84	0.2550	0.9703	0.8400	0.8268	0.1344	0.0521
No log	34.0	238	3.5466	0.825	0.2580	1.3118	0.825	0.8071	0.1553	0.0604
No log	35.0	245	3.5566	0.835	0.2574	1.1729	0.835	0.8134	0.1357	0.0592
No log	36.0	252	3.6022	0.83	0.2848	1.3337	0.83	0.8158	0.1528	0.0631
No log	37.0	259	3.5422	0.845	0.2598	1.2362	0.845	0.8311	0.1488	0.0588
No log	38.0	266	3.5677	0.825	0.2662	1.2509	0.825	0.8026	0.1417	0.0566
No log	39.0	273	3.5600	0.83	0.2673	1.2032	0.83	0.8128	0.1344	0.0561
No log	40.0	280	3.5818	0.82	0.2634	1.1062	0.82	0.8070	0.1389	0.0558
No log	41.0	287	3.5326	0.85	0.2520	1.3207	0.85	0.8404	0.1485	0.0529
No log	42.0	294	3.5954	0.83	0.2708	1.1103	0.83	0.8092	0.1401	0.0595
No log	43.0	301	3.5330	0.84	0.2528	1.2339	0.8400	0.8241	0.1405	0.0540
No log	44.0	308	3.5696	0.825	0.2654	1.1943	0.825	0.8079	0.1358	0.0544
No log	45.0	315	3.5438	0.83	0.2558	1.1267	0.83	0.8138	0.1314	0.0530
No log	46.0	322	3.5537	0.845	0.2497	1.2612	0.845	0.8298	0.1338	0.0525
No log	47.0	329	3.5609	0.85	0.2467	1.4284	0.85	0.8315	0.1333	0.0563
No log	48.0	336	3.5723	0.835	0.2595	1.1814	0.835	0.8187	0.1402	0.0545
No log	49.0	343	3.5591	0.825	0.2485	1.1736	0.825	0.8072	0.1429	0.0536
No log	50.0	350	3.5715	0.825	0.2585	1.3645	0.825	0.8098	0.1445	0.0564
No log	51.0	357	3.5813	0.83	0.2617	1.2375	0.83	0.8210	0.1371	0.0551
No log	52.0	364	3.6084	0.835	0.2592	1.2465	0.835	0.8168	0.1557	0.0550
No log	53.0	371	3.5574	0.84	0.2474	1.1932	0.8400	0.8255	0.1351	0.0543
No log	54.0	378	3.5863	0.85	0.2428	1.2885	0.85	0.8346	0.1347	0.0536
No log	55.0	385	3.5510	0.83	0.2520	1.2654	0.83	0.8163	0.1342	0.0543
No log	56.0	392	3.5516	0.835	0.2476	1.0430	0.835	0.8210	0.1336	0.0549
No log	57.0	399	3.5754	0.835	0.2475	1.3656	0.835	0.8245	0.1165	0.0528
No log	58.0	406	3.6017	0.83	0.2584	1.3561	0.83	0.8198	0.1490	0.0542
No log	59.0	413	3.5767	0.845	0.2488	1.2699	0.845	0.8357	0.1291	0.0521
No log	60.0	420	3.5844	0.835	0.2513	1.2919	0.835	0.8218	0.1286	0.0541
No log	61.0	427	3.5744	0.84	0.2443	1.2315	0.8400	0.8334	0.1441	0.0515
No log	62.0	434	3.5948	0.825	0.2505	1.2265	0.825	0.8052	0.1266	0.0539
No log	63.0	441	3.5833	0.845	0.2403	1.2410	0.845	0.8382	0.1268	0.0506
No log	64.0	448	3.6000	0.845	0.2451	1.2889	0.845	0.8282	0.1408	0.0526
No log	65.0	455	3.6050	0.84	0.2497	1.2870	0.8400	0.8298	0.1401	0.0534
No log	66.0	462	3.5950	0.86	0.2381	1.3004	0.8600	0.8491	0.1231	0.0511
No log	67.0	469	3.6030	0.85	0.2435	1.2246	0.85	0.8374	0.1252	0.0517
No log	68.0	476	3.6028	0.85	0.2433	1.2260	0.85	0.8404	0.1370	0.0513
No log	69.0	483	3.6112	0.86	0.2438	1.2148	0.8600	0.8487	0.1362	0.0522
No log	70.0	490	3.6147	0.85	0.2409	1.2230	0.85	0.8401	0.1258	0.0511
No log	71.0	497	3.6231	0.845	0.2391	1.2277	0.845	0.8341	0.1332	0.0506
3.5909	72.0	504	3.6298	0.845	0.2415	1.2897	0.845	0.8346	0.1288	0.0508
3.5909	73.0	511	3.6384	0.85	0.2427	1.2893	0.85	0.8401	0.1366	0.0515
3.5909	74.0	518	3.6364	0.845	0.2420	1.2224	0.845	0.8346	0.1219	0.0511
3.5909	75.0	525	3.6471	0.845	0.2441	1.2252	0.845	0.8346	0.1322	0.0517
3.5909	76.0	532	3.6469	0.85	0.2423	1.2259	0.85	0.8404	0.1300	0.0513
3.5909	77.0	539	3.6493	0.85	0.2423	1.2253	0.85	0.8401	0.1248	0.0514
3.5909	78.0	546	3.6534	0.85	0.2434	1.2273	0.85	0.8404	0.1271	0.0512
3.5909	79.0	553	3.6588	0.845	0.2430	1.2254	0.845	0.8346	0.1307	0.0513
3.5909	80.0	560	3.6636	0.845	0.2434	1.2249	0.845	0.8346	0.1259	0.0513
3.5909	81.0	567	3.6670	0.845	0.2433	1.2253	0.845	0.8346	0.1356	0.0513
3.5909	82.0	574	3.6689	0.845	0.2427	1.2256	0.845	0.8346	0.1365	0.0511
3.5909	83.0	581	3.6724	0.845	0.2433	1.2278	0.845	0.8346	0.1315	0.0510
3.5909	84.0	588	3.6768	0.85	0.2431	1.2260	0.85	0.8401	0.1374	0.0510
3.5909	85.0	595	3.6782	0.85	0.2424	1.2265	0.85	0.8401	0.1340	0.0509
3.5909	86.0	602	3.6817	0.85	0.2428	1.2261	0.85	0.8401	0.1332	0.0510
3.5909	87.0	609	3.6822	0.85	0.2427	1.2266	0.85	0.8401	0.1330	0.0508
3.5909	88.0	616	3.6835	0.85	0.2425	1.2259	0.85	0.8401	0.1328	0.0510
3.5909	89.0	623	3.6854	0.85	0.2425	1.2260	0.85	0.8401	0.1328	0.0509
3.5909	90.0	630	3.6874	0.85	0.2426	1.2259	0.85	0.8401	0.1327	0.0510
3.5909	91.0	637	3.6891	0.85	0.2428	1.2264	0.85	0.8401	0.1327	0.0510
3.5909	92.0	644	3.6903	0.85	0.2426	1.2265	0.85	0.8401	0.1328	0.0509
3.5909	93.0	651	3.6913	0.85	0.2427	1.2264	0.85	0.8401	0.1327	0.0509
3.5909	94.0	658	3.6922	0.85	0.2427	1.2265	0.85	0.8401	0.1326	0.0509
3.5909	95.0	665	3.6930	0.85	0.2426	1.2262	0.85	0.8401	0.1326	0.0510
3.5909	96.0	672	3.6936	0.85	0.2427	1.2266	0.85	0.8401	0.1327	0.0509
3.5909	97.0	679	3.6940	0.85	0.2426	1.2264	0.85	0.8401	0.1325	0.0510
3.5909	98.0	686	3.6946	0.85	0.2427	1.2265	0.85	0.8401	0.1326	0.0510
3.5909	99.0	693	3.6948	0.85	0.2427	1.2266	0.85	0.8401	0.1325	0.0510
3.5909	100.0	700	3.6948	0.85	0.2427	1.2265	0.85	0.8401	0.1325	0.0510

Framework versions

Transformers 4.26.1
Pytorch 1.13.1.post200
Datasets 2.9.0
Tokenizers 0.13.2

vit-tiny_tobacco3482_kd_NKD_t1.0_g1.5

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

NSDT 3DConvert

UnrealSynth

DreamTexture.js