README.md

metadata

license: apache-2.0
tags:
  - generated_from_trainer
datasets:
  - imagefolder
metrics:
  - accuracy
  - f1
  - recall
  - precision
model-index:
  - name: vit-base-patch16-224-in21k_male_or_female_eyes
    results:
      - task:
          name: Image Classification
          type: image-classification
        dataset:
          name: imagefolder
          type: imagefolder
          config: default
          split: train
          args: default
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.9726681127982646
          - name: F1
            type: f1
            value: 0.9741273100616017
          - name: Recall
            type: recall
            value: 0.9665851670741646
          - name: Precision
            type: precision
            value: 0.9817880794701986
language:
  - en
pipeline_tag: image-classification

vit-base-patch16-224-in21k_male_or_female_eyes

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: 0.0810
• Accuracy: 0.9727
• F1: 0.9741
• Recall: 0.9666
• Precision: 0.9818

Model description

This is a binary classification model to distinguish between male and female eyes.

For more information on how it was created, check out the following link: https://github.com/DunnBC22/Vision_Audio_and_Multimodal_Projects/blob/main/Computer%20Vision/Image%20Classification/Binary%20Classification/Male%20or%20Female%20Eyes/are_they_male_or_female_eyes_ViT.ipynb

Intended uses & limitations

This model is intended to demonstrate my ability to solve a complex problem using technology.

Training and evaluation data

Dataset Source: https://www.kaggle.com/datasets/pavelbiz/eyes-rtte

Sample Images From Dataset:

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: 3

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Recall	Precision
0.1998	1.0	577	0.2365	0.9072	0.9196	0.9976	0.8530
0.0846	2.0	1154	0.0810	0.9727	0.9741	0.9666	0.9818
0.0309	3.0	1731	0.0852	0.9809	0.9821	0.9837	0.9805

Framework versions

• Transformers 4.25.1
• Pytorch 1.12.1
• Datasets 2.8.0
• Tokenizers 0.12.1