finetuned-electrical-images

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

• Loss: 0.3726
• Accuracy: 0.8861

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: 0.0002
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.7116	0.4651	100	0.6399	0.7921
0.6953	0.9302	200	0.5589	0.8086
0.4078	1.3953	300	0.4946	0.8399
0.5852	1.8605	400	0.4872	0.8399
0.4993	2.3256	500	0.4687	0.8597
0.4479	2.7907	600	0.3986	0.8845
0.4101	3.2558	700	0.4385	0.8729
0.283	3.7209	800	0.4413	0.8762
0.3959	4.1860	900	0.4121	0.8729
0.318	4.6512	1000	0.4397	0.8696
0.2401	5.1163	1100	0.4887	0.8680
0.1273	5.5814	1200	0.4224	0.8663
0.1101	6.0465	1300	0.4378	0.8779
0.1773	6.5116	1400	0.3730	0.8845
0.2248	6.9767	1500	0.3726	0.8861
0.0987	7.4419	1600	0.4398	0.8845
0.16	7.9070	1700	0.4171	0.8828
0.1224	8.3721	1800	0.4336	0.8878
0.2111	8.8372	1900	0.3948	0.8944
0.112	9.3023	2000	0.4004	0.8944
0.0962	9.7674	2100	0.4092	0.8927

Framework versions

• Transformers 4.40.1
• Pytorch 2.2.1+cu121
• Datasets 2.19.0
• Tokenizers 0.19.1