SGL LLM Training Lab's Datasets generates synthetic LLM training data on demand from a prompt, for fine-tuning models. It supports open models on a private decentralized grid as well as frontier models, and can be used by humans or AI agents via x402.
