ibivibiv / giant-hydra-moe-240b

Text Generation238B10118Updated Feb 11, 2024

Downloads

101

Likes

18

Last updated

Feb 11, 2024

License

llama2

Yes, you read that correctly, this is a 4x70b MOE model with ~240B parameters. I doubt there is any way that I will have the benchmarks run here anytime soon to be on the leaderboard but I am looking into renting time on runpod to get the scores myself and put them here.

Tasks & library

  • text generation
  • transformers

Tags

  • transformers
  • mixtral
  • text-generation
  • moe
  • moerge
  • en

Technical specifications

Parameters238B
Librarytransformers
Languagesen
Licensellama2