Cerebras runs Llama 3.1 70B at 2,100 tokens per second, live demo available

(twitter.com)

6 points | by modeless 3 hours ago ago

1 comments

  • 3 hours ago
    [deleted]