Unity is Power: Semi-Asynchronous Collaborative Training of Large-Scale Models with Structured Pruning in Resource-Limited Clients

In this work, we study to release the potential of massive heterogeneous weak computing power to collaboratively train large-scale models on dispersed datasets. In order to improve both efficiency and accuracy in resource-adaptive collaborative learning, we take the first step to consider the $\mathit{\text{unstructured}}\mathit{\text{pruning}}$, $\mathit{\text{varying}}\mathit{\text{submodel}}\mathit{\text{architectures}}$, $\mathit{\text{knowledge}}\mathit{\text{loss}}$, and $\mathit{\text{straggler}}$ challenges simultaneously. We propose a novel semi-asynchronous collaborative training framework, namely ${Co\text{-}S}^{2}P$, with data distribution-aware structured pruning and cross-block knowledge transfer mechanism to address the above concerns. Furthermore, we provide theoretical proof that ${Co\text{-}S}^{2}P$ can achieve asymptotic optimal convergence rate of $O(1\mathrm{/}\sqrt{N^{\ast }EQ})$. Finally, we conduct extensive experiments on two types of tasks with a real-world hardware testbed including diverse IoT devices.The experimental results demonstrate that $Co\text{-}{S}^{2}P$ improves accuracy by up to 8.8% and resource utilization by up to 1.2$\times$ compared to state-of-the-art methods, while reducing memory consumption by approximately 22% and training time by about 24% on all resource-limited devices.