metadata

license: apache-2.0
tags:
  - generated_from_trainer
datasets:
  - imagefolder
metrics:
  - accuracy
  - f1
  - recall
  - precision
model-index:
  - name: vit-base-patch16-224-in21k_covid_19_ct_scans
    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: 0.94
          - name: F1
            type: f1
            value: 0.9379310344827586
          - name: Recall
            type: recall
            value: 0.8947368421052632
          - name: Precision
            type: precision
            value: 0.9855072463768116
language:
  - en
pipeline_tag: image-classification

vit-base-patch16-224-in21k_covid_19_ct_scans

This model is a fine-tuned version of google/vit-base-patch16-224-in21k.

It achieves the following results on the evaluation set:

Loss: 0.1727
Accuracy: 0.94
F1: 0.9379
Recall: 0.8947
Precision: 0.9855

Model description

This is a binary classification model to distinguish between CT scans that detect COVID-19 and those who do not.

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/COVID19%20Lung%20CT%20Scans/COVID19_Lung_CT_Scans_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/luisblanche/covidct

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
0.6742	1.0	38	0.4309	0.9	0.8993	0.8816	0.9178
0.6742	2.0	76	0.3739	0.8467	0.8686	1.0	0.7677
0.6742	3.0	114	0.1727	0.94	0.9379	0.8947	0.9855

Framework versions

Transformers 4.22.2
Pytorch 1.12.1
Datasets 2.5.2
Tokenizers 0.12.1