It seems likely that Groq models are not directly supported for evaluation in Phoenix, as they are primarily mentioned in the context of tracing. However, you might consider using a workaround by implementing a custom model class similar to how other custom models are integrated. For instance, you could create a custom class that interfaces with Groq models using a similar approach to the CustomModel example provided for other LLMs.Would you like more detail on how to implement such a custom model class? Sources:
- https://github.com/Arize-ai/phoenix/issues/2117
馃挕聽Hint: Mention RunLLM in the thread for followups.