Update app.py

app.py

@@ -1,23 +1,21 @@
 import gradio as gr
-from llama_cpp import Llama
-import os
+from ctransformers import AutoModelForCausalLM
 import subprocess
+import os
 
 MODEL = "unsloth.Q8_0.gguf"
 
-
 print("Downloading model...")
 subprocess.run([
     "wget", "-O", MODEL,
     "https://huggingface.co/Afifsudoers/NightPrompt_RV1_Instruct_8B_GGUF/resolve/main/unsloth.Q8_0.gguf?download=true"
 ], check=True)
 
-
-llm = Llama(
-    model_path=MODEL,
-    n_threads=8,      
-    n_ctx=2048,       
-    n_batch=128       
+llm = AutoModelForCausalLM(
+    MODEL,
+    model_type="llama",
+    n_threads=8,
+    max_new_tokens=256
 )
 
 def chat_fn(message, history):
@@ -26,14 +24,8 @@ def chat_fn(message, history):
         prompt += f"User: {user}\nAssistant: {assistant}\n"
     prompt += f"User: {message}\nAssistant:"
 
-    output = llm(
-        prompt,
-        max_tokens=256,
-        stop=["User:", "Assistant:"],
-        echo=False
-    )
-    text = output["choices"][0]["text"].strip()
-    return text
+    output = llm(prompt, stop=["User:", "Assistant:"])
+    return output
 
 demo = gr.ChatInterface(
     fn=chat_fn,