Train-automation

This an automation for computer vision classification training that help you to:

• preprocesse image with (rescale/resize).
• train model for wathever model architecture from tensorflow.keras.applications.

python train.py --data data/ --model_name model/model.h5 --size 128 --rescale --batch_size 64 --epochs 10

  --data: Directory of images to train
  --model_name: path to save model
  --size: size of Images
  --rescale: rescale the image

Test-automation

This an automation for computer vision classification testing that help you to:

• preprocesse image with (greyscale/rescale/resize)
• predict one image classe.
• evaluate the test set for the chosen model by multiple metrics(categorical crossentropy/F1 score/accuracy/recall/precision).

python auto_test_Dir.py --testdir Dir_test --model model.h5 --resize 128 --rescale

  --testdir: Directory of images to test
  --model: model path
  --resize: new size
  --rescale: rescale the image
  --greyscale: greyscale the image

MIT License