GitLab

use ctransformers as backup to llama.cpp

- change list to models
- remove generate from cli in favour of run

fix run arg parser

remove add cmd

fix imports

update generate to yield object

remove server extras for now