Zephyr7b gives gobbly gook output but Mistral7b works fine.

c0sogi / llama-api

An OpenAI-like LLaMA inference API

MIT License

111 stars 9 forks source link

Open Dougie777 opened 12 months ago

Dougie777 commented 12 months ago

Could there be some new format of gguf that we need to update the code for or something?