Simply make AI models faster, cheaper, smaller, greener!

Pruna AI makes, in one line of code, any AI model faster, cheaper, smaller, greener on any hardware. It covers CV, NLP, audio, graphs for predictive and generative AI.

We provide two packages. You can read their documentations to know more here.

• pruna: Package to smash your AI model to make it more efficient without losing quality. For this, you only need to call pruna.smash(). If you want to compress models on your side, you can request access here.
• pruna_engine: Package to run your AI model more efficiently without changing your pipeline. If you want to run models publicly shared on HuggingFace, you can install it from here.