Getting Started

• Create a new venv with python -m venv .venv
• Activate the venv with source ./.venv/bin/activate

Install Dependencies

• pip install -r requirements.txt

Train the model

This will take some time and output a model in the root directory model_state.pt

• python train.py

Test the model

You will find a line that looks like: img = Image.open('test_images/img_1.jpg') Just update the image to test the model against different images. The available images are:

• img_1.jpg (Should be a 2)
• img_2.jpg (Should be a 0)
• img_3.jpg (should be a 9)

To run the inferencing use:

• python inference.py

Documentation

• Pytorch
• ReLU
• Back Propagation
• ConvNet

Cuda

• install conda
• install cuda