You can optionally use different model backends with the --model flag.
- LLM:
gpt-4-1106-preview - VLM:
gpt-4-vision-preview - TTS:
tts-1 - STT:
whisper-1
pip install openai==1.1.1
export OPENAI_API_KEY=...
- LLM:
meta/llama-2-13b-chat - VLM:
yorickvp/llava-13b - TTS:
suno-ai/bark - STT:
openai/whisper
pip install replicate==0.20.0
export REPLICATE_API_TOKEN=...