INTELLECT-1 Technical Report

TL;DR

This paper presents INTELLECT-1, a 10-billion-parameter language model trained collaboratively across the globe using a novel distributed training framework, demonstrating the viability of decentralized, community-driven large-scale model training.

Contribution

It introduces PRIME, a scalable fault-tolerant distributed training framework, and demonstrates successful training of a large language model in a decentralized, community-driven setting.

Findings

Achieved 83-96% compute utilization during training.

Reduced communication bandwidth by 400x with custom all-reduce.

Trained a 10-billion-parameter model across global nodes.

Abstract

In this report, we introduce INTELLECT-1, the first 10 billion parameter language model collaboratively trained across the globe, demonstrating that large-scale model training is no longer confined to large corporations but can be achieved through a distributed, community-driven approach. INTELLECT-1 was trained on 1 trillion tokens using up to 14 concurrent nodes distributed across 3 continents, with contributions from 30 independent compute providers dynamically joining and leaving the training process, while maintaining 83-96% compute utilization and 36.2-41.4% model FLOPS utilization. We leverage PRIME, our scalable distributed training framework designed for fault-tolerant, high-performance training on unreliable, globally distributed nodes. Key innovations in PRIME include the ElasticDeviceMesh, which manages dynamic global process groups for fault-tolerant communication across the internet and local process groups for communication within a node, live checkpoint recovery kernels, and a hybrid DiLoCo-FSDP2 implementation. Using PRIME with DiLoCo and our custom int8 all-reduce, we achieve a 400x reduction in communication bandwidth compared to traditional data-parallel training settings while delivering comparable performance. These results demonstrate the feasibility and promise of training frontier foundation models in a decentralized network of global GPU resources.