18-tiny_tobacco3482_og_simkd

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: 269.3729
Accuracy: 0.36
Brier Loss: 0.8320
Nll: 4.3971
F1 Micro: 0.36
F1 Macro: 0.2725
Ece: 0.3539
Aurc: 0.6748

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: 16
eval_batch_size: 16
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	50	278.3639	0.075	0.8950	9.1030	0.075	0.0452	0.1624	0.8470
No log	2.0	100	276.7339	0.095	0.8890	8.6785	0.095	0.0820	0.1806	0.9095
No log	3.0	150	275.0460	0.16	0.8755	6.5857	0.16	0.0866	0.2296	0.8951
No log	4.0	200	274.4719	0.33	0.8662	6.6801	0.33	0.1980	0.3166	0.5931
No log	5.0	250	273.5627	0.23	0.8660	6.4504	0.23	0.1193	0.2606	0.7477
No log	6.0	300	273.5152	0.22	0.8622	6.1576	0.22	0.1323	0.2578	0.6395
No log	7.0	350	273.0874	0.2	0.8584	5.8986	0.2000	0.1249	0.2508	0.8129
No log	8.0	400	273.7433	0.19	0.8764	6.0516	0.19	0.1554	0.2550	0.8545
No log	9.0	450	272.4757	0.185	0.8533	5.9072	0.185	0.1340	0.2529	0.7360
273.9531	10.0	500	272.7899	0.14	0.8669	6.3132	0.14	0.1080	0.2373	0.8882
273.9531	11.0	550	271.7198	0.245	0.8484	4.7561	0.245	0.1370	0.2741	0.7203
273.9531	12.0	600	272.7391	0.095	0.8761	6.3636	0.095	0.0930	0.2120	0.9142
273.9531	13.0	650	271.8800	0.24	0.8566	4.9415	0.24	0.1926	0.2825	0.7850
273.9531	14.0	700	271.4349	0.305	0.8408	4.6971	0.305	0.1935	0.3087	0.7274
273.9531	15.0	750	271.7440	0.24	0.8476	3.9346	0.24	0.1835	0.2902	0.8098
273.9531	16.0	800	271.1944	0.28	0.8464	4.1470	0.28	0.2074	0.3074	0.7919
273.9531	17.0	850	271.2062	0.34	0.8363	4.8972	0.34	0.2136	0.3237	0.6403
273.9531	18.0	900	271.6591	0.295	0.8548	4.7423	0.295	0.1990	0.3066	0.7590
273.9531	19.0	950	271.5129	0.255	0.8571	4.8864	0.255	0.1898	0.2901	0.7929
270.6992	20.0	1000	271.4151	0.265	0.8543	5.2793	0.265	0.1924	0.2928	0.7842
270.6992	21.0	1050	271.0936	0.35	0.8449	4.4171	0.35	0.2347	0.3389	0.7230
270.6992	22.0	1100	271.2365	0.285	0.8504	4.7910	0.285	0.2020	0.3147	0.7840
270.6992	23.0	1150	271.3472	0.26	0.8496	4.1785	0.26	0.1800	0.2922	0.7983
270.6992	24.0	1200	271.1550	0.335	0.8466	4.8743	0.335	0.2518	0.3381	0.7148
270.6992	25.0	1250	270.7837	0.34	0.8455	4.7405	0.34	0.2392	0.3420	0.7592
270.6992	26.0	1300	271.2034	0.325	0.8485	5.0231	0.325	0.2418	0.3302	0.7411
270.6992	27.0	1350	270.5752	0.385	0.8334	4.5239	0.3850	0.2866	0.3535	0.6197
270.6992	28.0	1400	270.6892	0.35	0.8407	4.5651	0.35	0.2593	0.3428	0.6812
270.6992	29.0	1450	270.7246	0.36	0.8411	4.6054	0.36	0.2699	0.3365	0.6772
269.5806	30.0	1500	270.6962	0.32	0.8453	4.3997	0.32	0.2170	0.3145	0.7129
269.5806	31.0	1550	270.8489	0.325	0.8478	4.8399	0.325	0.2367	0.3180	0.6999
269.5806	32.0	1600	270.5093	0.36	0.8370	4.6472	0.36	0.2765	0.3456	0.6472
269.5806	33.0	1650	270.5440	0.35	0.8407	4.6948	0.35	0.2532	0.3292	0.6436
269.5806	34.0	1700	270.6743	0.395	0.8387	4.1379	0.395	0.2872	0.3730	0.6805
269.5806	35.0	1750	270.6282	0.33	0.8415	4.2499	0.33	0.2445	0.3410	0.7220
269.5806	36.0	1800	270.5709	0.31	0.8478	4.8275	0.31	0.2295	0.3063	0.7223
269.5806	37.0	1850	270.4933	0.33	0.8449	4.7300	0.33	0.2415	0.3157	0.6997
269.5806	38.0	1900	270.4860	0.355	0.8360	4.4343	0.3550	0.2598	0.3373	0.6545
269.5806	39.0	1950	269.9668	0.37	0.8306	4.4080	0.37	0.2602	0.3349	0.6056
268.8099	40.0	2000	270.3401	0.385	0.8389	5.0506	0.3850	0.2777	0.3567	0.6686
268.8099	41.0	2050	270.0377	0.365	0.8320	4.6238	0.3650	0.2466	0.3429	0.6167
268.8099	42.0	2100	270.2510	0.355	0.8363	4.7053	0.3550	0.2498	0.3352	0.6564
268.8099	43.0	2150	270.0252	0.38	0.8275	4.2261	0.38	0.2690	0.3557	0.5995
268.8099	44.0	2200	269.9756	0.36	0.8251	4.2336	0.36	0.2516	0.3226	0.6005
268.8099	45.0	2250	270.1610	0.34	0.8383	4.5052	0.34	0.2593	0.3164	0.6854
268.8099	46.0	2300	270.0320	0.33	0.8396	4.4659	0.33	0.2458	0.3114	0.7070
268.8099	47.0	2350	269.9535	0.335	0.8316	4.6707	0.335	0.2419	0.3203	0.6503
268.8099	48.0	2400	270.0645	0.365	0.8320	4.4182	0.3650	0.2517	0.3230	0.6360
268.8099	49.0	2450	270.0541	0.375	0.8348	4.4803	0.375	0.2717	0.3504	0.6444
268.2443	50.0	2500	269.8891	0.33	0.8374	4.7223	0.33	0.2365	0.3223	0.6716
268.2443	51.0	2550	269.8773	0.39	0.8355	4.4693	0.39	0.2637	0.3547	0.6525
268.2443	52.0	2600	269.6983	0.39	0.8298	4.5670	0.39	0.2716	0.3624	0.6230
268.2443	53.0	2650	270.0016	0.35	0.8327	4.3052	0.35	0.2402	0.3367	0.6602
268.2443	54.0	2700	269.7764	0.35	0.8318	4.3783	0.35	0.2432	0.3189	0.6373
268.2443	55.0	2750	269.7582	0.36	0.8272	4.4652	0.36	0.2535	0.3130	0.6197
268.2443	56.0	2800	269.7654	0.385	0.8253	4.4241	0.3850	0.2751	0.3375	0.6019
268.2443	57.0	2850	269.7094	0.375	0.8331	4.5019	0.375	0.2691	0.3413	0.6391
268.2443	58.0	2900	269.7442	0.37	0.8323	4.5422	0.37	0.2813	0.3311	0.6468
268.2443	59.0	2950	269.7930	0.385	0.8265	4.4502	0.3850	0.2820	0.3331	0.6032
267.7843	60.0	3000	269.6804	0.375	0.8270	4.3152	0.375	0.2723	0.3221	0.6005
267.7843	61.0	3050	269.7132	0.36	0.8355	4.4928	0.36	0.2625	0.3464	0.6679
267.7843	62.0	3100	269.6766	0.375	0.8308	4.5392	0.375	0.2644	0.3272	0.6307
267.7843	63.0	3150	269.5538	0.32	0.8345	4.4571	0.32	0.2299	0.3160	0.6791
267.7843	64.0	3200	269.5080	0.375	0.8262	4.1608	0.375	0.2581	0.3323	0.6357
267.7843	65.0	3250	269.6766	0.355	0.8362	4.4220	0.3550	0.2647	0.3250	0.6591
267.7843	66.0	3300	269.4889	0.38	0.8249	4.3636	0.38	0.2786	0.3452	0.6364
267.7843	67.0	3350	269.4171	0.4	0.8246	4.2888	0.4000	0.2877	0.3479	0.6114
267.7843	68.0	3400	269.6962	0.37	0.8286	4.3171	0.37	0.2614	0.3376	0.6369
267.7843	69.0	3450	269.3859	0.425	0.8225	4.2805	0.425	0.3053	0.3649	0.5751
267.4508	70.0	3500	269.5482	0.345	0.8377	4.7547	0.345	0.2559	0.3341	0.6728
267.4508	71.0	3550	269.3435	0.4	0.8216	4.3030	0.4000	0.2858	0.3391	0.6045
267.4508	72.0	3600	269.2284	0.405	0.8198	4.3223	0.405	0.3006	0.3598	0.5984
267.4508	73.0	3650	269.5614	0.37	0.8300	4.3579	0.37	0.2645	0.3515	0.6597
267.4508	74.0	3700	269.2826	0.39	0.8217	4.3413	0.39	0.2738	0.3317	0.5869
267.4508	75.0	3750	269.3375	0.39	0.8226	4.2796	0.39	0.2712	0.3485	0.6140
267.4508	76.0	3800	269.2377	0.39	0.8198	4.4825	0.39	0.2771	0.3450	0.5936
267.4508	77.0	3850	269.4238	0.375	0.8278	4.5048	0.375	0.2736	0.3349	0.6386
267.4508	78.0	3900	269.4522	0.4	0.8283	4.5498	0.4000	0.2870	0.3318	0.6187
267.4508	79.0	3950	269.4372	0.355	0.8316	4.5034	0.3550	0.2621	0.3182	0.6581
267.2267	80.0	4000	269.4058	0.38	0.8262	4.4096	0.38	0.2834	0.3453	0.6197
267.2267	81.0	4050	269.2977	0.37	0.8251	4.4666	0.37	0.2669	0.3374	0.6406
267.2267	82.0	4100	269.3194	0.37	0.8286	4.4947	0.37	0.2751	0.3332	0.6326
267.2267	83.0	4150	269.2936	0.365	0.8301	4.3184	0.3650	0.2725	0.3358	0.6815
267.2267	84.0	4200	269.2045	0.4	0.8242	4.2851	0.4000	0.2947	0.3526	0.6133
267.2267	85.0	4250	269.2916	0.38	0.8256	4.4965	0.38	0.2872	0.3190	0.6079
267.2267	86.0	4300	269.3817	0.35	0.8327	4.3865	0.35	0.2686	0.3341	0.6660
267.2267	87.0	4350	269.2809	0.36	0.8296	4.4521	0.36	0.2664	0.3367	0.6620
267.2267	88.0	4400	269.1650	0.405	0.8181	4.2997	0.405	0.2857	0.3527	0.5624
267.2267	89.0	4450	269.1071	0.375	0.8297	4.3559	0.375	0.2737	0.3386	0.6423
267.0481	90.0	4500	269.3120	0.385	0.8268	4.4102	0.3850	0.2833	0.3438	0.6264
267.0481	91.0	4550	269.1590	0.4	0.8243	4.2495	0.4000	0.2964	0.3382	0.5986
267.0481	92.0	4600	269.1963	0.4	0.8230	4.4469	0.4000	0.3070	0.3521	0.5969
267.0481	93.0	4650	269.1320	0.39	0.8256	4.2528	0.39	0.2856	0.3355	0.6072
267.0481	94.0	4700	269.2391	0.4	0.8222	4.3351	0.4000	0.2891	0.3392	0.6019
267.0481	95.0	4750	269.2017	0.395	0.8243	4.3717	0.395	0.2931	0.3435	0.6271
267.0481	96.0	4800	269.4085	0.37	0.8290	4.3941	0.37	0.2817	0.3255	0.6242
267.0481	97.0	4850	269.2195	0.4	0.8231	4.4454	0.4000	0.2889	0.3452	0.5916
267.0481	98.0	4900	269.3515	0.375	0.8296	4.3291	0.375	0.2754	0.3370	0.6380
267.0481	99.0	4950	269.3496	0.395	0.8244	4.4743	0.395	0.2989	0.3432	0.6100
266.9625	100.0	5000	269.3729	0.36	0.8320	4.3971	0.36	0.2725	0.3539	0.6748

Framework versions

Transformers 4.26.1
Pytorch 1.13.1.post200
Datasets 2.9.0
Tokenizers 0.13.2

jordyvl/18-tiny_tobacco3482_og_simkd

18-tiny_tobacco3482_og_simkd

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions