Explore

I'm trying to automate a service that is getting a picture of a client and returns the clothes he has on him, a coat for example. I tried to code a dedicated tensorflow based module that will provide this service, and his accuracy seems fine. The model is trained based on "fashion_mnist" dataset (https://www.kaggle.com/datasets/zalando-research/fashionmnist), and is working properly with the images that the dataset...

Originally posted 2023-09-06 Tagged: software_engineering Obligatory disclaimer: all opinions are mine and not of my employer As the joke goes, the two hardest problems in computer science are 1) cache invalidation and 2) naming. Like all good jokes, there’s a kernel of truth in there. Naming is poetry; the perfect ...

I am creating a machine learning model to classify images, and I am creating my datasets. I have a folder that contains my train, test, and validation datasets (train_ds, test_ds, val_ds respectively). I then define data augmentation as follows: tf.random.set_seed(42) data_augmentation = tf.keras.Sequential([ layers.RandomFlip("horizontal_and_vertical"), layers.experimental.preprocessing.RandomRotation(0.2), layers.experimental.preprocessing.RandomZoom( height_factor=(-0.3, -0.03), width_factor=None), ]) and I change my train_ds to account for data augmentation as follows. This code...