Update app.py

app.py

@@ -1,4 +1,4 @@
-import streamlit as st
+import gradio as gr
 from llama_cpp import Llama
 import os
 import requests
@@ -15,14 +15,14 @@ def download_model(url, save_path):
                 file.write(chunk)
         return "Model downloaded successfully."
     except Exception as e:
-        return f"Error downloading model: {e}"
+        return f"Error: {e}"
 
 def load_model(model_path):
     try:
-        llm = Llama(model_path, n_threads=2, n_gpu_layers=0)
+        llm = Llama(model_path, n_threads=2, n_gpu_layers=0) #force cpu, and set threads.
         return llm
     except Exception as e:
-        return f"Error loading model: {e}"
+        return f"Error: {e}"
 
 def apply_chat_template(model_name, messages, system_prompt):
     model_name_lower = model_name.lower()
@@ -53,19 +53,17 @@ def generate_response(prompt, model):
         output = model(prompt, max_tokens=256)
         return output["choices"][0]["text"].strip()
     except Exception as e:
-        return f"Error generating response: {e}"
+        return f"Error: {e}"
 
 def inference(message, history, model_url, system_prompt):
     if model_url and not os.path.exists(MODEL_PATH):
         download_result = download_model(model_url, MODEL_PATH)
         if "Error" in download_result:
-            st.session_state.messages.append({"role": "assistant", "content": download_result})
-            return
+            return history + [{"role": "assistant", "content": download_result}], history
 
     llm = load_model(MODEL_PATH)
     if isinstance(llm, str):
-        st.session_state.messages.append({"role": "assistant", "content": llm})
-        return
+        return history + [{"role": "assistant", "content": llm}], history
 
     messages = [{"role": "system", "content": system_prompt}]
     for item in history:
@@ -74,25 +72,18 @@ def inference(message, history, model_url, system_prompt):
 
     prompt = apply_chat_template(llm.model_path, messages, system_prompt)
     response = generate_response(prompt, llm)
-    st.session_state.messages.append({"role": "assistant", "content": response})
+    history.append({"role": "assistant", "content": response})
+    return history, history
 
-st.title("llama.cpp Chat")
+with gr.Blocks() as iface:
+    model_url_input = gr.Textbox(label="Model URL (GGUF)", placeholder="Enter GGUF model URL...")
+    system_prompt_input = gr.Textbox(label="System Prompt", value=DEFAULT_SYSTEM_PROMPT, lines=3)
+    chatbot = gr.Chatbot(type="messages")
+    message = gr.Textbox(label="Message")
+    send_button = gr.Button("Send")
+    state = gr.State([])
 
-if "messages" not in st.session_state:
-    st.session_state.messages = []
+    send_button.click(inference, inputs=[message, state, model_url_input, system_prompt_input], outputs=[chatbot, state])
+    message.submit(inference, inputs=[message, state, model_url_input, system_prompt_input], outputs=[chatbot, state])
 
-model_url = st.text_input("Model URL (GGUF)", placeholder="Enter GGUF model URL...")
-system_prompt = st.text_area("System Prompt", value=DEFAULT_SYSTEM_PROMPT)
-
-for message in st.session_state.messages:
-    with st.chat_message(message["role"]):
-        st.markdown(message["content"])
-
-if prompt := st.chat_input("Message"):
-    st.session_state.messages.append({"role": "user", "content": prompt})
-    with st.chat_message("user"):
-        st.markdown(prompt)
-    history = [m for m in st.session_state.messages if m["role"] != "system"]
-    inference(prompt, history, model_url, system_prompt)
-    with st.chat_message("assistant"):
-        st.markdown(st.session_state.messages[-1]["content"])
+iface.launch()