Seventy-billion-parameter Llama 3 model for text generation and instruction following.
Meta released the 70B Llama 3 pretrained and instruction-tuned models.
