NicoHambauer / teachable-machine-student-contest

Student contest in training the best CV classifier with https://teachablemachine.withgoogle.com/train/image

Teachable Machine Student Contest Environment Setup

This guide provides instructions on how to set up the Python environment required for running the Teachable Machine project on macOS, specifically optimized for Apple Silicon (M1/M2 chips).

Requirements

• Conda (Miniconda or Anaconda)
• macOS with Apple Silicon (M1/M2)

Conda Environment Installation with YAML File

To use this YAML file, save it as teachable-ml.yml and then run the following command to create the Conda environment:

conda env create -f teachable-ml.yml

After creating the environment, activate it with:

conda activate teachable-ml

This setup ensures that users can quickly get started with the Teachable Machine project by following clear, step-by-step instructions, and using a YAML file reduces the potential for errors during the installation of dependencies.

Manual Environment Setup

1. Create a Conda Environment

   Open your terminal and run the following command to create a new Conda environment named teachable-ml with Python 3.9:

   conda create --name teachable-ml python=3.9

2. Activate the Environment

   Activate the newly created environment:

   conda activate teachable-ml

3. Install Required Packages

   Install tensorflow-macos, tensorflow-metal for GPU support, keras version 2.13.1, and numpy version 1.24.3 using pip:

   pip install tensorflow-macos tensorflow-metal pillow matplotlib scikit-learn keras==2.13.1 numpy==1.24.3

Setup Verification

After installation, verify the setup by running the following commands in your terminal:

python -c "import tensorflow as tf; print(tf.__version__)"
python -c "import keras; print(keras.__version__)"
python -c "import numpy; print(numpy.__version__)"
python -c "import sklearn; print(sklearn.__version__)"
python -c "import matplotlib; print(matplotlib.__version__)"
python -c "import PIL; print(PIL.__version__)"

If the installations were successful, you should see the versions of TensorFlow, Keras, and NumPy printed without errors.