Fixed

app.py

@@ -3,7 +3,7 @@ import subprocess
 import sys
 import os
 
-# === RUNTIME INSTALL OF llama-cpp-python (fixes all HF issues) ===
+# === RUNTIME INSTALL OF llama-cpp-python ===
 try:
     from llama_cpp import Llama
     print("llama-cpp-python already installed.")
@@ -15,30 +15,30 @@ except ImportError:
             "https://github.com/yownas/llama-cpp-python-wheels/releases/download/v0.3.16/llama_cpp_python-0.3.16+cpuavx-cp310-cp310-linux_x86_64.whl"
         ])
         print("llama-cpp-python installed from wheel.")
-    except Exception as e:  # <-- fixed: added "as e" so it doesn't crash
-        print("Wheel failed → falling back to PyPI (2–4 min)...")
+    except Exception as e:
+        print("Wheel failed → falling back to PyPI...")
         subprocess.check_call([
             sys.executable, "-m", "pip", "install", "--no-cache-dir",
             "llama-cpp-python==0.3.16", "--force-reinstall"
         ])
-    from llama_cpp import Llama  # <-- fixed: must be inside the except block!
+    from llama_cpp import Llama  # ← INDENTED CORRECTLY
 
 from huggingface_hub import hf_hub_download
 
-# === YOUR FINE-TUNED LLAMA 3.2 3B ===
-MODEL_REPO = "your-username/your-model-repo"           # CHANGE THIS
-MODEL_FILE = "your-finetuned-llama-3.2-3b-q4_k_m.gguf"  # CHANGE THIS
+# === WORKING PUBLIC MODEL ===
+MODEL_REPO = "TheBloke/Llama-3.2-3B-Instruct-GGUF"
+MODEL_FILE = "llama-3.2-3b-instruct-q4_k_m.gguf"
 
-print("Downloading your fine-tuned Llama 3.2 3B model...")
+print("Downloading Llama 3.2 3B Instruct (Q4_K_M)...")
 model_path = hf_hub_download(
     repo_id=MODEL_REPO,
     filename=MODEL_FILE,
     local_dir="./models",
     local_dir_use_symlinks=False
 )
-print(f"Model ready: {model_path}")
+print(f"Model downloaded: {model_path}")
 
-print("Loading model (Llama 3.2 3B)...")
+print("Loading model into memory (20–40 sec)...")
 llm = Llama(
     model_path=model_path,
     n_ctx=8192,
@@ -47,7 +47,7 @@ llm = Llama(
     n_gpu_layers=0,
     verbose=False
 )
-print("Model loaded!")
+print("Model loaded — ready to chat!")
 
 def chat(message, history):
     if not message.strip():
@@ -65,7 +65,7 @@ def chat(message, history):
         max_tokens=512,
         temperature=0.7,
         top_p=0.9,
-        stop=["<|eot_id|>", "<|end_of_text|>"],  # <-- Llama 3.2 stop tokens
+        stop=["<|eot_id|>", "<|end_of_text|>"],
         stream=False
     )
 
@@ -73,7 +73,7 @@ def chat(message, history):
     history.append((message, bot_response))
     return history, ""
 
-# === CSS & INTERFACE (unchanged, perfect) ===
+# === CSS & UI (perfect) ===
 custom_css = """
 @import url('https://fonts.googleapis.com/css2?family=JetBrains+Mono:wght@400;700&family=Source+Code+Pro:wght@400;600&display=swap');
 body, .gradio-container { background: #0c0c0c !important; font-family: 'JetBrains Mono', monospace !important; }
@@ -86,38 +86,19 @@ button { background: #1a1a1a !important; border: 1px solid #00ff00 !important; c
 button:hover { background: #00ff00 !important; color: #000 !important; }
 .primary { background: #00ff00 !important; color: #000 !important; }
 footer { display: none !important; }
-::-webkit-scrollbar { width: 8px; background: #0c0c0c; }
-::-webkit-scrollbar-thumb { background: #00ff00; }
 """
 
 with gr.Blocks(theme=gr.themes.Base(primary_hue="green"), css=custom_css, title="$ LLAMA TERMINAL") as demo:
-    gr.Markdown("# $ LLAMA TERMINAL\n```\n> System Online | Neural Network Active\n> Type your query below...\n```")
+    gr.Markdown("# $ LLAMA TERMINAL\n```\n> System Online | Llama 3.2 3B Ready\n> Type your query below...\n```")
     chatbot = gr.Chatbot(height=600)
-
     with gr.Row():
         msg = gr.Textbox(placeholder="$ Enter command...", show_label=False, scale=8, container=False)
         submit = gr.Button("SEND", scale=1, variant="primary")
-
-    gr.Examples(
-        examples=[
-            "What is the capital of France?",
-            "Explain quantum computing",
-            "Write fibonacci in Python",
-            "Write a haiku about AI",
-        ],
-        inputs=msg
-    )
-
+    gr.Examples(["What is the capital of France?", "Write a haiku about AI"], inputs=msg)
     gr.ClearButton([msg, chatbot], value="CLEAR")
-
     submit.click(chat, [msg, chatbot], [chatbot, msg])
     msg.submit(chat, [msg, chatbot], [chatbot, msg])
 
 if __name__ == "__main__":
     demo.queue(max_size=20)
-    demo.launch(
-        share=True,
-        server_name="0.0.0.0",
-        server_port=7860,
-        show_error=True
-    )
+    demo.launch(server_name="0.0.0.0", server_port=7860, share=True, show_error=True)