Hello, thank you so much for letting me find this. A great project :) and potentially what I've been looking for for a long time. It offers the possibility to change many settings and potentially expand them.
Unfortunately, I'm having trouble getting it to run locally.
I installed it under Windows 11/WSL and finally got it to start.
- The API keys were replaced with "1"
In summary, the log only reads "error, repeat"
(openai.APIConnectionError: Connection error)
-
I tried to access an ollama server (also installed under WSL and loaded the DataSpecialist-v.1)
-
I tried to access a koboldcpp server under Windows (with a different model)
-
The log still says: "Warning: pyfiglet not installed. Install with 'pip install pyfiglet' for formatted banners." But it is installed.
I'm also surprised that /chat/completions is being addressed during execution, even though only
http://0.0.0.0:eg (8082) /v1
is being addressed.
Can someone guide me/us through the initial steps of implementing it locally?
I would be very grateful and could delve deeper into the actual project.
** I dont want to train a model yet. Just create and work with a dataset from different sources, so i set the "do_train: False"
Hello, thank you so much for letting me find this. A great project :) and potentially what I've been looking for for a long time. It offers the possibility to change many settings and potentially expand them.
Unfortunately, I'm having trouble getting it to run locally.
I installed it under Windows 11/WSL and finally got it to start.
In summary, the log only reads "error, repeat"
(openai.APIConnectionError: Connection error)
I tried to access an ollama server (also installed under WSL and loaded the DataSpecialist-v.1)
I tried to access a koboldcpp server under Windows (with a different model)
The log still says: "Warning: pyfiglet not installed. Install with 'pip install pyfiglet' for formatted banners." But it is installed.
I'm also surprised that /chat/completions is being addressed during execution, even though only
http://0.0.0.0:eg (8082) /v1
is being addressed.
Can someone guide me/us through the initial steps of implementing it locally?
I would be very grateful and could delve deeper into the actual project.
** I dont want to train a model yet. Just create and work with a dataset from different sources, so i set the "do_train: False"