vit-base-patch16-224-Trial006-YEL_STEM4

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: 0.0984
• 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
0.7587	0.57	1	0.6296	0.7111
0.6884	1.71	3	0.7730	0.4
0.6112	2.86	5	0.5194	0.8222
0.5566	4.0	7	0.6194	0.6444
0.5216	4.57	8	0.5428	0.6889
0.4488	5.71	10	0.3884	0.8222
0.4438	6.86	12	0.3301	0.8444
0.3897	8.0	14	0.2362	0.9111
0.3789	8.57	15	0.1942	0.9333
0.3484	9.71	17	0.3995	0.8222
0.2727	10.86	19	0.1636	0.9556
0.209	12.0	21	0.1489	0.9556
0.2253	12.57	22	0.1712	0.9111
0.2407	13.71	24	0.2239	0.9111
0.1615	14.86	26	0.0984	1.0
0.1735	16.0	28	0.1231	0.9111
0.179	16.57	29	0.1203	0.9111
0.1464	17.71	31	0.0422	1.0
0.1444	18.86	33	0.0409	1.0
0.1758	20.0	35	0.0394	1.0
0.199	20.57	36	0.0246	1.0
0.1525	21.71	38	0.0179	1.0
0.1536	22.86	40	0.0441	1.0
0.115	24.0	42	0.0836	0.9333
0.106	24.57	43	0.0654	0.9778
0.1267	25.71	45	0.0285	1.0
0.1264	26.86	47	0.0199	1.0
0.1554	28.0	49	0.0192	1.0
0.1456	28.57	50	0.0195	1.0

Framework versions

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