Update app.py

app.py

@@ -3,19 +3,24 @@ from llama_cpp import Llama
 import os
 from huggingface_hub import hf_hub_download
 
-# Configuration
-MODEL_REPO = "TheBloke/Mistral-7B-Instruct-v0.3-GGUF"
-MODEL_FILE = "mistral-7b-instruct-v0.3.Q4_K_M.gguf"
-MAX_TOKENS = 150
-CPU_THREADS = os.cpu_count() or 4  # Use all available cores
+# Verified working repositories (as of July 2024)
+MODEL_REPO = "bartowski/Mistral-7B-Instruct-v0.3-GGUF"
+MODEL_FILE = "Mistral-7B-Instruct-v0.3-Q4_K_M.gguf"
 
-# Download model (will cache after first run)
+# Download model
 model_path = hf_hub_download(
     repo_id=MODEL_REPO,
     filename=MODEL_FILE,
-    cache_dir="models"
+    token=os.environ.get("HF_TOKEN")  # Required for gated models
 )
 
+
+# Configuration
+MODEL_REPO = "TheBloke/Mistral-7B-Instruct-v0.3-GGUF"
+MODEL_FILE = "mistral-7b-instruct-v0.3.Q4_K_M.gguf"
+MAX_TOKENS = 150
+CPU_THREADS = os.cpu_count() 
+
 # Initialize Llama
 llm = Llama(
     model_path=model_path,