update commit with phi-3 mini 1114

app.py

@@ -1,26 +1,22 @@
 import gradio as gr
-import torch
 from transformers import AutoTokenizer, AutoModelForCausalLM, pipeline
 
-# ✅ Force ZeroGPU to allocate GPU early
+# ✅ Safe import of the GPU decorator
 try:
-    _ = torch.tensor([1.0], device="cuda")
-    print("✅ ZeroGPU triggered successfully.")
-except Exception as e:
-    print(f"⚠️ GPU allocation failed: {e}")
+    from spaces import GPU
+except ImportError:
+    def GPU(func): return func  # Fallback if not in a HF Space
 
-# 🚀 Load Phi-3 Mini model
+# ✅ Load Phi-3 Mini model
 model_id = "microsoft/phi-3-mini-4k-instruct"
 tokenizer = AutoTokenizer.from_pretrained(model_id)
 model = AutoModelForCausalLM.from_pretrained(
-    model_id,
-    torch_dtype=torch.float16 if torch.cuda.is_available() else torch.float32,
-    device_map="auto"
+    model_id, torch_dtype="auto", device_map="auto"
 )
-
 pipe = pipeline("text-generation", model=model, tokenizer=tokenizer)
 
-# 💬 Chat function
+# ✅ Apply GPU decorator to ensure ZeroGPU allocates GPU
+@GPU
 def chat_fn(message, history):
     history_text = ""
     for item in history:
@@ -33,18 +29,14 @@ def chat_fn(message, history):
     result = pipe(prompt, max_new_tokens=512, do_sample=True, temperature=0.7)[0]['generated_text']
     reply = result.split("<|assistant|>")[-1].strip()
 
-    # Format code blocks if applicable
-    if "```" not in reply and any(w in reply for w in ["def ", "class ", "import "]):
+    if "```" not in reply and any(word in reply for word in ["def ", "class ", "import "]):
         reply = f"```\n{reply}\n```"
-
     return reply
 
-# 🧩 Gradio UI
+# ✅ Gradio UI
 with gr.Blocks(theme=gr.themes.Soft()) as demo:
     gr.Markdown("## 💬 Chat with Phi-3 Mini")
-    gr.Markdown(
-        "Welcome to your AI Assistant powered by Phi-3 Mini and ZeroGPU (uses GPU if available)."
-    )
+    gr.Markdown("Welcome to your AI Assistant powered by Phi-3 Mini. Ask me anything or request code examples!")
 
     gr.ChatInterface(
         fn=chat_fn,
@@ -57,5 +49,4 @@ with gr.Blocks(theme=gr.themes.Soft()) as demo:
         chatbot=gr.Chatbot(type="messages")
     )
 
-# 🚀 Launch
-demo.launch(ssr_mode=False)
+demo.launch()