output_dir

This model is a fine-tuned version of google/vit-base-patch16-224-in21k on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 1.2976
• Accuracy: 0.6

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: 7e-05
• train_batch_size: 64
• eval_batch_size: 64
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 256
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: reduce_lr_on_plateau
• num_epochs: 77

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.8	2	2.0706	0.15
No log	2.0	5	2.0309	0.2313
No log	2.8	7	1.9846	0.2562
1.9868	4.0	10	1.8915	0.4062
1.9868	4.8	12	1.8529	0.3125
1.9868	6.0	15	1.7422	0.4125
1.9868	6.8	17	1.6761	0.4313
1.6815	8.0	20	1.6310	0.4562
1.6815	8.8	22	1.5900	0.45
1.6815	10.0	25	1.5402	0.4313
1.6815	10.8	27	1.5018	0.5
1.4233	12.0	30	1.4620	0.4875
1.4233	12.8	32	1.4286	0.5062
1.4233	14.0	35	1.4045	0.5125
1.4233	14.8	37	1.3860	0.5312
1.2127	16.0	40	1.3571	0.5
1.2127	16.8	42	1.3293	0.5375
1.2127	18.0	45	1.3742	0.4813
1.2127	18.8	47	1.3151	0.5437
1.0075	20.0	50	1.3053	0.5312
1.0075	20.8	52	1.3266	0.5375
1.0075	22.0	55	1.2964	0.5312
1.0075	22.8	57	1.2278	0.5875
0.8232	24.0	60	1.2501	0.5563
0.8232	24.8	62	1.2330	0.575
0.8232	26.0	65	1.2198	0.5625
0.8232	26.8	67	1.2071	0.5875
0.6738	28.0	70	1.2643	0.5875
0.6738	28.8	72	1.2594	0.5563
0.6738	30.0	75	1.2263	0.5312
0.6738	30.8	77	1.3218	0.5188
0.5715	32.0	80	1.2593	0.5312
0.5715	32.8	82	1.2214	0.5625
0.5715	34.0	85	1.3060	0.55
0.5715	34.8	87	1.2727	0.5563
0.4523	36.0	90	1.2749	0.5375
0.4523	36.8	92	1.3570	0.5437
0.4523	38.0	95	1.2815	0.5687
0.4523	38.8	97	1.2233	0.6062
0.3971	40.0	100	1.2097	0.6
0.3971	40.8	102	1.2881	0.5813
0.3971	42.0	105	1.2400	0.575
0.3971	42.8	107	1.3140	0.5375
0.3616	44.0	110	1.1525	0.6125
0.3616	44.8	112	1.2725	0.5938
0.3616	46.0	115	1.2634	0.5813
0.3616	46.8	117	1.2299	0.6
0.338	48.0	120	1.3408	0.5375
0.338	48.8	122	1.1931	0.5938
0.338	50.0	125	1.2806	0.5938
0.338	50.8	127	1.2410	0.575
0.3445	52.0	130	1.2901	0.5813
0.3445	52.8	132	1.2504	0.6062
0.3445	54.0	135	1.1614	0.5875
0.3445	54.8	137	1.2247	0.6062
0.3299	56.0	140	1.2591	0.5625
0.3299	56.8	142	1.2629	0.5687
0.3299	58.0	145	1.2369	0.5938
0.3299	58.8	147	1.2771	0.575
0.3292	60.0	150	1.3284	0.5875
0.3292	60.8	152	1.2550	0.5625
0.3292	61.6	154	1.3047	0.55

Framework versions

• Transformers 4.33.2
• Pytorch 2.0.1+cu118
• Datasets 2.14.5
• Tokenizers 0.13.3