vit-base-patch16-224-Trial007-YEL_STEM3

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

• Loss: 0.0840
• Accuracy: 1.0

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: 5e-05
• train_batch_size: 60
• eval_batch_size: 60
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 240
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.7133	0.89	2	0.6860	0.5926
0.6666	1.78	4	0.6151	0.7037
0.524	2.67	6	0.4915	0.8704
0.3759	4.0	9	0.3304	0.9074
0.3218	4.89	11	0.2177	0.9259
0.2764	5.78	13	0.1735	0.9630
0.2558	6.67	15	0.0840	1.0
0.7446	8.0	18	0.1256	0.9815
0.1251	8.89	20	0.0527	1.0
0.1491	9.78	22	0.0366	1.0
0.1559	10.67	24	0.0225	1.0
0.0916	12.0	27	0.0139	1.0
0.0751	12.89	29	0.0112	1.0
0.084	13.78	31	0.0102	1.0
0.0741	14.67	33	0.0103	1.0
0.065	16.0	36	0.0071	1.0
0.1019	16.89	38	0.0061	1.0
0.061	17.78	40	0.0052	1.0
0.0751	18.67	42	0.0045	1.0
0.0336	20.0	45	0.0064	1.0
0.0362	20.89	47	0.0038	1.0
0.073	21.78	49	0.0037	1.0
0.0748	22.67	51	0.0050	1.0
0.0543	24.0	54	0.0056	1.0
0.0408	24.89	56	0.0048	1.0
0.0729	25.78	58	0.0050	1.0
0.0638	26.67	60	0.0035	1.0
0.042	28.0	63	0.0029	1.0
0.0982	28.89	65	0.0025	1.0
0.0238	29.78	67	0.0023	1.0
0.0536	30.67	69	0.0032	1.0
0.1131	32.0	72	0.0029	1.0
0.0569	32.89	74	0.0030	1.0
0.0717	33.78	76	0.0034	1.0
0.0567	34.67	78	0.0037	1.0
0.0807	36.0	81	0.0039	1.0
0.0888	36.89	83	0.0039	1.0
0.0548	37.78	85	0.0039	1.0
0.078	38.67	87	0.0040	1.0
0.0453	40.0	90	0.0041	1.0
0.0442	40.89	92	0.0042	1.0
0.0425	41.78	94	0.0041	1.0
0.0619	42.67	96	0.0041	1.0
0.0437	44.0	99	0.0040	1.0
0.0552	44.44	100	0.0040	1.0

Framework versions

• Transformers 4.30.0.dev0
• Pytorch 1.12.1
• Datasets 2.12.0
• Tokenizers 0.13.1