Create app.py

app.py

@@ -0,0 +1,50 @@
+import gradio as gr
+from unsloth import FastLanguageModel
+
+# 1. Load your model (This takes a minute on first run)
+# We use 4-bit loading to fit it on smaller hardware
+model, tokenizer = FastLanguageModel.from_pretrained(
+    model_name = "SALEETAI/coding-agent-qwen-sft", # Your Model ID
+    max_seq_length = 2048,
+    load_in_4bit = True,
+)
+FastLanguageModel.for_inference(model)
+
+# 2. Define the Chat Function
+def chat_with_model(message, history):
+    # Format the conversation (Qwen format)
+    messages = []
+    for user_msg, bot_msg in history:
+        messages.append({"role": "user", "content": user_msg})
+        messages.append({"role": "assistant", "content": bot_msg})
+    messages.append({"role": "user", "content": message})
+    
+    # Tokenize and Generate
+    inputs = tokenizer.apply_chat_template(
+        messages, 
+        tokenize=True, 
+        add_generation_prompt=True, 
+        return_tensors="pt"
+    )
+    
+    # Generate response (Streamed)
+    outputs = model.generate(
+        input_ids=inputs, 
+        max_new_tokens=512, 
+        use_cache=True
+    )
+    
+    # Decode just the new tokens
+    response = tokenizer.decode(outputs[0][len(inputs[0]):], skip_special_tokens=True)
+    return response
+
+# 3. Build the Interface
+demo = gr.ChatInterface(
+    fn=chat_with_model,
+    title="Code Agent Qwen-7B (SFT)",
+    description="Ask me to write Python, SQL, or C++ code!",
+    examples=["Write a Python function to reverse a string.", "Fix this bug: for i in list: list.remove(i)"],
+)
+
+if __name__ == "__main__":
+    demo.launch()