AllenAI has released Olmo-core 3, an open-source training infrastructure tailored specifically for large Mixture of Experts (MoE) models. By open-sourcing the complex scaffolding required to wrangle these sprawling architectures, they are inviting the broader community to shoulder the computational misery of frontier model training. * Olmo-core 3 provides open scalable training

Hugging Face has introduced the Open TTS Leaderboard, a scalable platform designed to evaluate multilingual text-to-speech and voice cloning models. Because the world was clearly crying out for yet another benchmark to obsess over, this initiative promises to bring some much-needed order to the chaotic landscape of synthetic voice generation.