3 ms·
I just tested 120B from the Groq API on agentic stuff (multi-step function calling, similar to claude code) and it's not that good. Agentic fine-tuning seems ke
by parhamn 1y ago
I just tested 120B from the Groq API on agentic stuff (multi-step function calling, similar to claude code) and it's not that good. Agentic fine-tuning seems key, hopefully someone drops one soon.
- AmazingTurtle 1y agoIm not sure if groq uses the proper harmony template?