Update app.py

app.py

@@ -1,4 +1,4 @@
-import gradio as gr
+"""import gradio as gr
 import torch
 from transformers import AutoTokenizer, AutoModelForCausalLM
 import spaces
@@ -61,4 +61,66 @@ gradio_app = gr.ChatInterface(
 )
 
 if __name__ == "__main__":
-    gradio_app.launch()
+    gradio_app.launch()"""
+
+
+import gradio as gr
+import torch
+from transformers import AutoTokenizer, AutoModelForCausalLM
+
+model_name = "hosseinhimself/ISANG-v1.0-8B"
+
+# Load tokenizer globally
+tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)
+
+def load_model():
+    try:
+        # Load the model on CPU
+        model = AutoModelForCausalLM.from_pretrained(
+            model_name,
+            torch_dtype=torch.float32,  # Use float32 for CPU
+            trust_remote_code=True,
+            low_cpu_mem_usage=True
+        )
+        model.to("cpu")  # Explicitly set the model to CPU
+        print("Model loaded successfully on CPU.")
+        return model
+    except Exception as e:
+        print(f"Error loading model: {e}")
+        raise
+
+def chat(prompt, history):
+    model = load_model()
+    # Add system prompt
+    system_prompt = "You are ISANG, a multilingual large language model made by ISANG AI. You only respond in Persian, Korean, or English. If a user uses one of these languages, reply in the same language."
+
+    # Format the conversation history with system prompt
+    context = f"System: {system_prompt}\n"
+    for user_message, bot_message in history:
+        context += f"User: {user_message}\nBot: {bot_message}\n"
+    context += f"User: {prompt}\nBot:"
+
+    # Generate a response
+    inputs = tokenizer(context, return_tensors="pt", truncation=True, max_length=512)
+    with torch.no_grad():
+        outputs = model.generate(**inputs, max_new_tokens=200, temperature=0.7)
+    response = tokenizer.decode(outputs[0], skip_special_tokens=True)
+
+    # Extract the latest response
+    response = response[len(context):].strip()
+    history.append((prompt, response))
+    return history, response
+
+gradio_app = gr.ChatInterface(
+    fn=chat,
+    title="ISANG Chatbot",
+    description="This is a chatbot powered by the ISANG model. Enter your messages to chat with it!",
+    examples=[
+        ["سلام، چطوری؟"],
+        ["برام یه داستان تعریف کن"],
+        ["نظرت درباره هوش مصنوعی چیه؟"]
+    ]
+)
+
+if __name__ == "__main__":
+    gradio_app.launch()