senko
commented
4 months ago Try prefixing the model name with openai/, eg openai/meta-llama/Llama-3-8b-chat-hf.
Also please note that we've just released version 0.2, which changes configuration format (it's now a JSON file) which allows (hopefully!) easier setup of local LLMs and even using different ones for different agents.
The extension that supports it is currently in pre-release mode and we plan to make a full update to the extension this week.
Version
VisualStudio Code extension
Operating System
Ubuntu Linux
What happened?
Cannot get gpt pilot to work with together.ai, getting this error: Response text: {"error":{"message":"Unable to access model Llama-3-8b-chat-hf. Please visit https://api.together.xyz/ to see the list of supported models or contact the owner to request access.","type":"invalid_request_error","param":null,"code":"model_not_found"}}
I'm not sure if its removing the prefix from the model I add in the .env: MODEL_NAME=meta-llama/Llama-3-8b-chat-hf