This model is a fine-tuned version of google/vit-large-patch32-384 on the imagefolder dataset. It achieves the following results on the evaluation set:

Loss: 0.1157
Accuracy: 0.9756

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: 8
eval_batch_size: 8
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 3
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.3336	0.03	100	0.2980	0.9325
0.0235	0.07	200	0.1580	0.9612
0.0381	0.1	300	0.2212	0.9540
0.0507	0.14	400	0.4664	0.9037
0.0052	0.17	500	0.1737	0.9670
0.0499	0.21	600	0.2187	0.9511
0.0454	0.24	700	0.1837	0.9569
0.0317	0.28	800	0.2616	0.9497
0.0594	0.31	900	0.1867	0.9555
0.0583	0.35	1000	0.1817	0.9569
0.0044	0.38	1100	0.2358	0.9497
0.0836	0.42	1200	0.2422	0.9454
0.0712	0.45	1300	0.1943	0.9555
0.0399	0.49	1400	0.2922	0.9440
0.0098	0.52	1500	0.3783	0.9325
0.0414	0.56	1600	0.2583	0.9454
0.1085	0.59	1700	0.2241	0.9511
0.0492	0.63	1800	0.2813	0.9368
0.044	0.66	1900	0.3361	0.9353
0.0344	0.7	2000	0.2549	0.9468
0.002	0.73	2100	0.1794	0.9641
0.0731	0.77	2200	0.2300	0.9540
0.0151	0.8	2300	0.2050	0.9569
0.0031	0.84	2400	0.2175	0.9454
0.1015	0.87	2500	0.1725	0.9626
0.0383	0.91	2600	0.2104	0.9540
0.0926	0.94	2700	0.1762	0.9540
0.0001	0.98	2800	0.1978	0.9612
0.1365	1.01	2900	0.1512	0.9655
0.083	1.04	3000	0.1298	0.9641
0.0002	1.08	3100	0.1976	0.9540
0.0042	1.11	3200	0.1719	0.9698
0.0002	1.15	3300	0.1924	0.9583
0.0002	1.18	3400	0.1732	0.9626
0.0978	1.22	3500	0.1902	0.9612
0.1067	1.25	3600	0.1868	0.9612
0.0005	1.29	3700	0.2166	0.9468
0.0007	1.32	3800	0.2293	0.9425
0.0001	1.36	3900	0.2296	0.9626
0.0001	1.39	4000	0.1685	0.9684
0.0001	1.43	4100	0.2106	0.9655
0.0004	1.46	4200	0.1614	0.9670
0.0	1.5	4300	0.1311	0.9727
0.0	1.53	4400	0.1445	0.9784
0.0433	1.57	4500	0.1544	0.9727
0.0263	1.6	4600	0.2133	0.9626
0.0	1.64	4700	0.1903	0.9598
0.0	1.67	4800	0.1587	0.9583
0.0	1.71	4900	0.1817	0.9655
0.1503	1.74	5000	0.2346	0.9526
0.0699	1.78	5100	0.1143	0.9713
0.0004	1.81	5200	0.1937	0.9626
0.0001	1.85	5300	0.2660	0.9540
0.2208	1.88	5400	0.1500	0.9713
0.0494	1.92	5500	0.1203	0.9698
0.0001	1.95	5600	0.1231	0.9756
0.0001	1.99	5700	0.1254	0.9698
0.0	2.02	5800	0.1622	0.9684
0.0001	2.06	5900	0.1464	0.9698
0.0	2.09	6000	0.1420	0.9698
0.0	2.12	6100	0.1416	0.9698
0.0	2.16	6200	0.1408	0.9698
0.0001	2.19	6300	0.1402	0.9698
0.0147	2.23	6400	0.1536	0.9655
0.0	2.26	6500	0.1944	0.9612
0.0	2.3	6600	0.1724	0.9684
0.0003	2.33	6700	0.1910	0.9612
0.0003	2.37	6800	0.1995	0.9626
0.0004	2.4	6900	0.1563	0.9655
0.0	2.44	7000	0.1460	0.9727
0.0	2.47	7100	0.1434	0.9727
0.0	2.51	7200	0.1242	0.9741
0.0041	2.54	7300	0.1364	0.9713
0.0	2.58	7400	0.1396	0.9684
0.0	2.61	7500	0.1371	0.9655
0.0	2.65	7600	0.1373	0.9684
0.0	2.68	7700	0.1230	0.9698
0.0	2.72	7800	0.1225	0.9698
0.0	2.75	7900	0.1223	0.9698
0.0001	2.79	8000	0.1218	0.9698
0.0	2.82	8100	0.1186	0.9756
0.0	2.86	8200	0.1183	0.9756
0.0	2.89	8300	0.1167	0.9756
0.0	2.93	8400	0.1163	0.9756
0.0	2.96	8500	0.1162	0.9756
0.0	3.0	8600	0.1157	0.9756

Framework versions

Transformers 4.39.3
Pytorch 2.1.2
Datasets 2.18.0
Tokenizers 0.15.2

ZaneHorrible/ViTL-32-384-1e4-batch_16_epoch_4_classes_24

ViTL-32-384-1e4-batch_16_epoch_4_classes_24

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

ZaneHorrible/
ViTL-32-384-1e4-batch_16_epoch_4_classes_24