4 pointsby faangguyindia8 hours ago2 comments
  • ramigb7 hours ago
    That's awesome I came from the jev in python thread. is the model shareable? do you have any more writing on this? would love to read more about it! cool demo anyways.
  • faangguyindia8 hours ago
    Here you can see a 26B model hacked using a trained KV cache bank, responding like Gemma. The latency I am seeing is <127 ms.