Week 2 journal

My (team's) last week goal:

- Done image data preprocessing, generate my dataset from image files in a directory.

- Start researching model creation

My work toward goal:

- I created "tf.keras.preprocessing.image_dataset_from_directory" to load data from google drive and I will have train data, validation data and test data.

- Labels are the class names of my dataset. Validation data is 0.1 split of train dataset.

- Randomly printed out image and label from dataset to check my data.

- Created training model, and got the best model is with 907,092 parameters, which has 84% evaluate accuracy with test data.

My (team's) next week goal:

- Use fewer parameters and find better models.