I'm just wondering about the exact version of Llama used for the plugin and if the model is being upgraded regularly. Is it still using the 2.x or has it been upgraded to 3.x or even 4? I believe the devs are actively working behind the scene and it would be very helpful if there's a bit of additional notes in the plugin page such as current version, changelogs, etc.

1 comment

sorted by: hot top controversial new old
[–] 1 point 1 year ago

The tokenizer used for the tokenizing is Llama-2 Based. So, I'm assuming it is a Llama 2 base, but not sure if the model used is the same as the tokenizer. This is the most recent update from the dev regarding the AI models.

  • source