Moshi /moshiko /moshika speech text foundation LLM by KyutAI

Hi, KyutAI finally released their moshi model. It's a "speech text foundational model", kinda like the promised gpt-4o but it outputs text at the same time as speech. Here's their link: https://github.com/kyutai-labs/moshi Here's the hf link: https://huggingface.co/collections/kyutai/moshi-v01-release-66eaeaf3302bef6bd9ad7acd It exists in several version and some are already quantized