Welcome to DFQ-Dojo

We are dedicated to advancing model quantization research and providing high-performance, efficient AI models. Our focus is on data-free, zero-shot quantization, enabling models to run faster and with smaller memory footprints without requiring access to the original training data.

Our Mission

Efficiency without compromise: Make state-of-the-art models lighter, faster, and deployable on resource-constrained devices.

Data privacy first: Achieve quantization without relying on the original training data, protecting sensitive datasets.

Open collaboration: Share models, tools, and research with the AI community to accelerate adoption and innovation.