🧀 BigCheese.ai

Social

Llama 3.1 Omni Model

🧀

LLaMA-Omni is a cutting-edge speech interaction model that provides low-latency, high-quality end-to-end speech capabilities. Building on Llama-3.1-8B-Instruct, it aims for performance at the GPT-4 level, featuring simultaneous generation of text and speech responses and can be trained in less than 3 days with just 4 GPUs.

  • Built on Llama-3.1
  • Supported by 4 GPUs
  • Low latency of 226ms
  • Generates text and speech
  • Apache-2.0 license