Add input id model

app.py

@@ -30,17 +30,17 @@ def load_model(model_id):
 if not torch.cuda.is_available():
     DESCRIPTION += "\n<p>Running on CPU 🥶 This demo does not work on CPU.</p>"
 
-if torch.cuda.is_available():
-    model_id = "Nekochu/Luminia-13B-v3"
-    model, tokenizer = load_model(model_id)
-
+MODELS = [
+    "Nekochu/Luminia-13B-v3",
+    "Nekochu/Llama-2-13B-German-ORPO",
+]
 
 @spaces.GPU(duration=120)
 def generate(
+    model_id: str,
     message: str,
     chat_history: list[tuple[str, str]],
     system_prompt: str,
-    model_id: str = "Nekochu/Luminia-13B-v3",
     max_new_tokens: int = 1024,
     temperature: float = 0.6,
     top_p: float = 0.9,
@@ -48,6 +48,7 @@ def generate(
     repetition_penalty: float = 1.2,
 ) -> Iterator[str]:
     model, tokenizer = load_model(model_id)
+
     conversation = []
     if system_prompt:
         conversation.append({"role": "system", "content": system_prompt})
@@ -81,12 +82,14 @@ def generate(
         outputs.append(text)
         yield "".join(outputs)
 
+# Combine predefined models into a single placeholder for the textbox
+model_placeholder = "\n".join(MODELS)
 
 chat_interface = gr.ChatInterface(
     fn=generate,
     additional_inputs=[
+        gr.Textbox(label="Select Model or Enter Custom Model ID", placeholder=model_placeholder),
         gr.Textbox(label="System prompt", lines=6),
-        gr.Textbox(label="Model ID", placeholder="Nekochu/Luminia-13B-v3"),
         gr.Slider(
             label="Max new tokens",
             minimum=1,
@@ -137,4 +140,4 @@ with gr.Blocks(css="style.css") as demo:
     gr.Markdown(LICENSE)
 
 if __name__ == "__main__":
-    demo.queue(max_size=20).launch()
+    demo.queue(max_size=20).launch()