This repository contains code for Efficient image loading blog post.

Tested with Python 3.12.x.

python3 -m venv .venv
source .venv/bin/activate
pip install -U pip
pip install -r requirements.txt

Example benchmark run:

python benchmark.py --path images --method cv2 pil lmdb tfrecords --mode RGB --iters 10

The benchmark now creates the LMDB output directory automatically and filters directory inputs down to real image files before loading them.

Want to become an expert in AI? AI Courses by OpenCV is a great place to start.