attempt9 fix

app.py

@@ -1,7 +1,5 @@
-import os
-from threading import Thread
+rom threading import Thread
 from typing import Iterator
-
 import gradio as gr
 import spaces
 import torch
@@ -11,33 +9,27 @@ MAX_MAX_NEW_TOKENS = 2048
 DEFAULT_MAX_NEW_TOKENS = 1024
 MAX_INPUT_TOKEN_LENGTH = int(os.getenv("MAX_INPUT_TOKEN_LENGTH", "4096"))
 
+MODELS = {
+    "Nekochu/Luminia-13B-v3": "Default - Nekochu/Luminia-13B-v3",
+    "Nekochu/Llama-2-13B-German-ORPO": "German ORPO - Nekochu/Llama-2-13B-German-ORPO",
+}
+
 DESCRIPTION = """\
-# Nekochu/Luminia-13B-v3
-This Space demonstrates model [Nekochu/Luminia-13B-v3](https://huggingface.co/Nekochu/Luminia-13B-v3) by Nekochu, a Llama 2 model with 13B parameters fine-tuned for SD gen prompt 
-"""
+# Text Generation with Selectable Models
 
-LICENSE = """
-<p/>
----.
+This Space demonstrates text generation using different models. Choose a model from the dropdown and experience its creative capabilities!
 """
 
-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 = AutoModelForCausalLM.from_pretrained(model_id, device_map="auto", load_in_4bit=True)
-    tokenizer = AutoTokenizer.from_pretrained(model_id)
-    tokenizer.use_default_system_prompt = False
+LICENSE = """<p/> ---."""
 
+if not torch.cuda.is_available():
+    DESCRIPTION += "\n<p>Running on CPU  This demo does not work on CPU.</p>"
 
-@spaces.GPU(duration=120)
 def generate(
-    model_id: str,
     message: str,
     chat_history: list[tuple[str, str]],
     system_prompt: str,
+    model_id: str,
     max_new_tokens: int = 1024,
     temperature: float = 0.6,
     top_p: float = 0.9,
@@ -80,11 +72,11 @@ def generate(
         outputs.append(text)
         yield "".join(outputs)
 
-
+model_dropdown = gr.Dropdown(label="Select Model", choices=list(MODELS.values()))
 chat_interface = gr.ChatInterface(
     fn=generate,
     additional_inputs=[
-        gr.Dropdown(label="Model ID", choices=["Nekochu/Luminia-13B-v3", "Nekochu/Llama-2-13B-German-ORPO"]),
+        model_dropdown,
         gr.Textbox(label="System prompt", lines=6),
         gr.Slider(
             label="Max new tokens",
@@ -123,10 +115,9 @@ chat_interface = gr.ChatInterface(
         ),
     ],
     stop_btn=None,
-    outputs="text",
     examples=[
-        ["Nekochu/Luminia-13B-v3", "### Instruction: Create stable diffusion metadata based on the given english description. Luminia ### Input: favorites and popular SFW ### Response:"],
-        ["Nekochu/Llama-2-13B-German-ORPO", "### Instruction: Provide tips on stable diffusion to optimize low token prompts and enhance quality include prompt example. ### Response:"],
+        ["### Instruction: Create stable diffusion metadata based on the given english description. Luminia ### Input: favorites and popular SFW ### Response:"],
+        ["### Instruction: Provide tips on stable diffusion to optimize low token prompts and enhance quality include prompt example. ### Response:"],
     ],
 )
 
@@ -137,4 +128,4 @@ 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()