3 ms·

My personal experience is that Deepseek far exceeds code llama of the same size, but it was released quite a while ago.

• turnsout 3y ago
  Agreed—I hope Meta studied Deepseek's approach. The idea of a Deepseek Coder at 70B would be exciting.
  • whimsicalism 3y ago
    The “approach” is likely just training on more tokens.
    • imjonse 3y ago
      It's the quality of data and the method of training, not just the amount of tokens (per their paper release a few days ago)
  • hackerlight 3y ago
    There's a deepseek coder around 30-35b and it has almost identical performance to the 7b on benchmarks.