I tried using the proxy recently and encountered a few issues, this was pointing it at a local llama.cpp
first item noticed was trying to set DEBUG logging didn't work, AI said it was being overridden in another module.
second the server target set in the proxy config is ignored on load so if you don't also set the default url thing to the same as the proxy address it won't fetch a model list.
and then when using crush it gave some content error and caused some crash output on terminal.
I had AI fix the crash locally and it then gave some other error crush side so something is messed up beyond just that crash.
I didn't save any of the logs or modifications since I got annoyed and it was just a venv pip install.
oh and the documentation needs updating as there seems to be several versions of how things work laying around.
I tried using the proxy recently and encountered a few issues, this was pointing it at a local llama.cpp
first item noticed was trying to set DEBUG logging didn't work, AI said it was being overridden in another module.
second the server target set in the proxy config is ignored on load so if you don't also set the default url thing to the same as the proxy address it won't fetch a model list.
and then when using crush it gave some content error and caused some crash output on terminal.
I had AI fix the crash locally and it then gave some other error crush side so something is messed up beyond just that crash.
I didn't save any of the logs or modifications since I got annoyed and it was just a venv pip install.
oh and the documentation needs updating as there seems to be several versions of how things work laying around.