Switch to wizard-vicuna-uncensored:13b + optimized persona

- Changed all model references to wizard-vicuna-uncensored:13b
- Updated persona.txt for better character alignment
- Config updated in character_genesis.json
- Core engines updated (NaturalLanguageEngine, NaturalLanguageEngine_LLM)
- FastAPI runtime updated
This commit is contained in:
Arch Agent
2026-03-10 16:08:30 +01:00
parent 80eec32e22
commit 63e6e2ded0
5 changed files with 27 additions and 41 deletions

View File

@@ -15,7 +15,7 @@ PATHS = {
# Admin nutzt Llama für intensive Antworten, normale User Mistral
MODEL_NORMAL = "mistral-nemo:latest"
MODEL_ADMIN = "llama3-8b-abliterated:latest"
MODEL_ADMIN = "wizard-vicuna-uncensored:13b"
def load_admin_config():
try: