Paris: A Decentralized Trained Open-Weight Diffusion Model
This addresses the challenge of resource-intensive training for large-scale diffusion models, making it more accessible by eliminating the need for synchronized GPU clusters, though it is incremental as it builds on existing decentralized training concepts.
The paper tackles the problem of training high-quality text-to-image diffusion models without centralized infrastructure by introducing Paris, a decentralized model that partitions data into clusters for independent expert training, achieving generation quality comparable to centralized baselines while using 14× less training data and 16× less compute than prior decentralized methods.
We present Paris, the first publicly released diffusion model pre-trained entirely through decentralized computation. Paris demonstrates that high-quality text-to-image generation can be achieved without centrally coordinated infrastructure. Paris is open for research and commercial use. Paris required implementing our Distributed Diffusion Training framework from scratch. The model consists of 8 expert diffusion models (129M-605M parameters each) trained in complete isolation with no gradient, parameter, or intermediate activation synchronization. Rather than requiring synchronized gradient updates across thousands of GPUs, we partition data into semantically coherent clusters where each expert independently optimizes its subset while collectively approximating the full distribution. A lightweight transformer router dynamically selects appropriate experts at inference, achieving generation quality comparable to centrally coordinated baselines. Eliminating synchronization enables training on heterogeneous hardware without specialized interconnects. Empirical validation confirms that Paris's decentralized training maintains generation quality while removing the dedicated GPU cluster requirement for large-scale diffusion models. Paris achieves this using 14$\times$ less training data and 16$\times$ less compute than the prior decentralized baseline.