Project code for Udacity's AI Programming with Python Nanodegree program. In this project, students first develop code for an image classifier built with PyTorch, then convert it into a command line application.
Considering that the program should be ran in either cpu or gpu, some of the parts in the notebook might be out of order, I consider it to be an exploratoty phase. Towards the end of Part 2, I was able to enable Cuda on my personal laptop gpu and speed up the process of training a model.
A few things to consider going through the notebook:
- Every part on the rubric has been labeled in code segments
- Although some parts aren't being modularized in this implementation, a more concise approach is done on the second part
Personally ran in RTX GeForce 3060 Laptop
Torch version: conda install pytorch torchvision torchaudio pytorch-cuda=12.1 -c pytorch -c nvidia
Basic usage: python train.py data_directory
Prints out training loss, validation loss, and validation accuracy as the network trains
Options:
- Set directory to save checkpoints:
python train.py data_dir --save_dir save_directory - Choose architecture:
python train.py data_dir --arch "densenet121" - Set hyperparameters:
python train.py data_dir --learning_rate 0.01 --hidden_units 512 --epochs 20 - Use GPU for training:
python train.py data_dir --gpu gpu
Basic usage: python predict.py /path/to/image checkpoint
Options:
- Return top K most likely classes:
python predict.py input checkpoint --top_k 3 - Use a mapping of categories to real names:
python predict.py input checkpoint --category_names cat_to_name.json - Use GPU for inference:
python predict.py input checkpoint --gpu