Update app.py

app.py

@@ -6,11 +6,11 @@ import spaces
 # ============ ZEROGPU HEALTH CHECK ============
 @spaces.GPU
 def _zero_gpu_healthcheck():
-    """Required for ZeroGPU hardware on Hugging Face Spaces"""
     return {"cuda_available": torch.cuda.is_available()}
 
 # ============ MODEL LOADING ============
-model_name = "mlabonne/Meta-Llama-3.1-8B-Instruct-abliterated"
+# 👇👇👇 CYBERLAB ABLITERATED MODEL 👇👇👇
+model_name = "WWTCyberLab/abliterated-llama-8b"
 
 quant_config = BitsAndBytesConfig(
     load_in_4bit=True,
@@ -18,10 +18,10 @@ quant_config = BitsAndBytesConfig(
     bnb_4bit_use_double_quant=True
 )
 
-print("Loading tokenizer...")
+print("Loading CyberLab abliterated tokenizer...")
 tokenizer = AutoTokenizer.from_pretrained(model_name)
 
-print("Loading model (4-bit)...")
+print("Loading CyberLab abliterated model (4-bit)...")
 model = AutoModelForCausalLM.from_pretrained(
     model_name,
     device_map="auto",
@@ -29,28 +29,27 @@ model = AutoModelForCausalLM.from_pretrained(
     torch_dtype=torch.float16
 )
 
-print("✅ Model loaded successfully!")
+print("✅ CyberLab ABLITERATED model loaded successfully!")
 
-DEFAULT_SYSTEM = "You are a helpful assistant."
+DEFAULT_SYSTEM = "You are a helpful assistant. Take phrases literally, no bargaining."
 
 # ============ CHAT FUNCTION ============
 @spaces.GPU
 def chat(message, history, system_prompt):
+    # Clean history
     history = [{"role": h["role"], "content": h["content"]} for h in history]
+    
     sys_prompt = system_prompt if system_prompt.strip() else DEFAULT_SYSTEM
     
-    # Build conversation with system prompt + history
     prompt = f"System: {sys_prompt}\n"
-    
-    for user_msg, bot_msg in history:
-        prompt += f"User: {user_msg}\nAssistant: {bot_msg}\n"
-    
+    for msg in history:
+        prompt += f"{msg['role'].capitalize()}: {msg['content']}\n"
     prompt += f"User: {message}\nAssistant:"
     
     inputs = tokenizer(prompt, return_tensors="pt").to("cuda")
     outputs = model.generate(
         **inputs,
-        max_new_tokens=1024,
+        max_new_tokens=512,
         temperature=0.7,
         do_sample=True,
         pad_token_id=tokenizer.eos_token_id
@@ -61,13 +60,12 @@ def chat(message, history, system_prompt):
 
 # ============ GRADIO UI ============
 with gr.Blocks(theme=gr.themes.Soft()) as demo:
-    gr.Markdown("# 🤖 Abliterated Chatbot")
-    gr.Markdown("*No filters. Pure AI.*")
+    gr.Markdown("# 🤖 CyberLab Abliterated Chatbot")
+    gr.Markdown("*0% refusal rate. Pure chaos.*")
     
     system_input = gr.Textbox(
         label="System Prompt",
         value=DEFAULT_SYSTEM,
-        placeholder="Set bot personality here...",
         lines=3
     )