Petals: Run 100B+ language models at home bit-torrent style

Re:Petals: Run 100B+ language models at home bit-torrent style by sva_ on Monday January 02, 2023 @11:26AM (#34219738)

I tried the chat on http://chat.petals.ml, and it seems to struggle with the current load (as per the disclaimer at the top)

    Human: How is the weather today?
    
    AI: the
    AI theAI)aultAIAI ) course )
    . can?esterday to people?
    ? is to think thatified )

Really cool project though, I wanted to work on something similar.

[ Parent | on HN ]

Re:Petals: Run 100B+ language models at home bit-torrent style by throwaway743 on Monday January 02, 2023 @12:29PM (#34220714)

Won't even load for me atm

[ Parent | on HN ]

Re:Petals: Run 100B+ language models at home bit-torrent style by metadat on Monday January 02, 2023 @01:42PM (#34221834)

I replied to me with:

  It is nice today.

Not garbled, but also extremely shallow.

[ Parent | on HN ]

Re:Petals: Run 100B+ language models at home bit-torrent style by jck on Monday January 02, 2023 @02:48PM (#34222720)

To be fair, so was the question.

This is a language model, not an oracle or an interface to weather forecast data.

[ Parent | on HN ]

Re:Petals: Run 100B+ language models at home bit-torrent style by borzunov on Monday January 02, 2023 @03:21PM (#34223115)

It varies from time to time. You can also switch to the few-shot mode to try machine translation, code generation, or other tasks involving longer responses

[ Parent | on HN ]