Fashion-Mnist Classifier with PyTorch
Inference
1- clone this repository:
git clone https://github.com/Jhamed7/Fashion-Mnist-Classifier.git
2- Install required packages:
pip install -r requirements.txt
3-Run below command in terminal:
python inference.py
- by default this command uses test.png images to test classifier. to test other images(grayscale 28*28 pixels image), use below command:
python inference.py --image <path_to_image>
- also 'cpu' with use to run this command, use
--device cuda
to change this behavior.
Train
to train again model on Fashion-Mnist dataset, below command is the way:
python train.py --device <cpu or cuda> --batch_size <int> --epochs <int> --lr <learning rate: float>
this command assumes that you have dataset in a folder in a same directory.