Patch Groq preview models maxed to 8K tokens due to warning

This commit is contained in:
timothycarambat 2024-08-01 09:24:57 -07:00
parent 187f6ed9e6
commit 6dc3642661

View file

@ -49,7 +49,7 @@ class GroqLLM {
return 8192; return 8192;
case "llama-3.1-70b-versatile": case "llama-3.1-70b-versatile":
case "llama-3.1-8b-instant": case "llama-3.1-8b-instant":
return 131072; return 8000;
case "mixtral-8x7b-32768": case "mixtral-8x7b-32768":
return 32768; return 32768;
default: default: