Update modules/config_settings_public.py

modules/config_settings_public.py

@@ -1,41 +1,34 @@
-# FILE: modules/config_settings_public.py (Hugging Face Demo - v2.0 - Verified Models)
+# FILE: modules/config_settings_public.py (HF Demo - v8.0)
 
 import os
 import logging
 from huggingface_hub import hf_hub_download
 
-logger = logging.getLogger("ZOTHEOS_Config")
-if not logger.handlers:
-    handler = logging.StreamHandler()
-    formatter = logging.Formatter('%(asctime)s - %(name)s - %(levelname)s - [%(funcName)s] - %(message)s')
-    handler.setFormatter(formatter)
-    logger.addHandler(handler)
-    logger.setLevel(logging.INFO)
+logger = logging.getLogger("ZOTHEOS_Config_HF")
 
-# --- ✅ VERIFIED WEB-OPTIMIZED MODEL SOURCES ---
-# These models are smaller, faster.
+# --- ✅ DEFINITIVE & VERIFIED WEB-OPTIMIZED MODEL SOURCES ---
 MODEL_DEFINITIONS = {
     "mistral": {
-        "repo_id": "TheBloke/Mistral-7B-Instruct-v0.1-GGUF",
-        "filename": "mistral-7b-instruct-v0.2.Q2_K.gguf" # Smallest quantization for speed
+        "repo_id": "TheBloke/Mistral-7B-Instruct-v0.2-GGUF",
+        "filename": "mistral-7b-instruct-v0.2.Q2_K.gguf"
     },
     "gemma": {
-        "repo_id": "google/gemma-7b-it.gguf", # Using the official Google repository
-        "filename": "gemma-7b-it.gguf"
+        "repo_id": "TheBloke/gemma-2b-it-GGUF",
+        "filename": "gemma-2b-it.Q2_K.gguf"
     },
     "qwen": {
-        "repo_id": "second-state/Qwen1.5-1.8B-Chat-Q2_K.gguf",
-        "filename": "Qwen1.5-1.8B-Chat-Q2_K.gguf"
+        "repo_id": "Qwen/Qwen1.5-0.5B-Chat-GGUF",
+        "filename": "qwen1_5-0.5b-chat-q2_k.gguf"
     }
 }
 
 MODEL_PATHS = {}
 
-logger.info("✅✅✅ RUNNING IN WEB DEMO MODE (Hugging Face Space) ✅✅✅")
+logger.info("✅✅✅ RUNNING IN WEB DEMO MODE (True Fusion - CPU Survival) ✅✅✅")
 N_GPU_LAYERS_FALLBACK = 0  # Force CPU-only mode
 
 for name, model_info in MODEL_DEFINITIONS.items():
-    logger.info(f"Downloading model for demo: {name} from {model_info['repo_id']}")
+    logger.info(f"Downloading demo model: {name}...")
     try:
         MODEL_PATHS[name] = hf_hub_download(repo_id=model_info["repo_id"], filename=model_info["filename"])
         logger.info(f"✅ Successfully downloaded {name}")
@@ -43,12 +36,13 @@ for name, model_info in MODEL_DEFINITIONS.items():
         logger.error(f"❌ FAILED to download {name}: {e}")
         raise e
 
-# --- WEB-OPTIMIZED MODEL PARAMETERS ---
+# --- ✅ CPU-OPTIMIZED MODEL PARAMETERS ---
 MODEL_SPECIFIC_PARAMS = {
     "_default": { 
         "n_gpu_layers": N_GPU_LAYERS_FALLBACK,
-        "n_ctx": 2048, # Smaller context for lower RAM
-        "n_batch": 512,
+        "n_ctx": 1024,
+        "n_batch": 256,
+        "n_threads": 4,
         "verbose": True
     },
     "mistral": { "chat_format": "mistral-instruct" },
@@ -56,15 +50,15 @@ MODEL_SPECIFIC_PARAMS = {
     "qwen": { "chat_format": "chatml" }
 }
 
-# --- AGI-TIER INFERENCE & PROMPTS ---
-INFERENCE_PRESETS = {"balanced": {"temperature": 0.7, "max_tokens": 512}}
+# --- TIER INFERENCE & PROMPTS ---
+INFERENCE_PRESETS = {"balanced": {"temperature": 0.7, "max_tokens": 256}}
 DEFAULT_INFERENCE_PRESET = "balanced"
-DEFAULT_SYSTEM_PROMPT = "You are ZOTHEOS, an ethical AI. Respond only in English."
+DEFAULT_SYSTEM_PROMPT = "You are ZOTHEOS, an ethical AI. Respond concisely and only in English."
 MODEL_ROLES = {"mistral": "analyst", "gemma": "humanist", "qwen": "skeptic"}
 MODEL_ROLE_SYSTEM_PROMPTS = {
-    "analyst": "You are an impartial analyst. Provide structured, logical insights. Respond only in English.",
-    "humanist": "You are an empathetic AI. Focus on the emotional and ethical impact. Respond only in English.",
-    "skeptic": "You are a respectful skeptic. Question assumptions and highlight risks. Respond only in English.",
+    "analyst": "You are an analyst. Be logical. Respond only in English.",
+    "humanist": "You are a humanist. Focus on values. Respond only in English.",
+    "skeptic": "You are a skeptic. Challenge the premise. Respond only in English.",
 }
 
-logger.info("✅ Hugging Face Demo Configuration Loaded Successfully.")
+logger.info("✅ Hugging Face Demo (CPU Survival Mode) Configuration Loaded.")