Better description

app.py

@@ -140,13 +140,14 @@ monospace_css = """
 
 css += share_btn_css + monospace_css + ".gradio-container {color: black}"
 
-description = """
+description = f"""
 <div style="text-align: center;">
     <h1> Phind VS WizardCoder Playground</h1>
 </div>
 <div style="text-align: left;">
-    <p>This is a demo to generate text and code with the following <a href="https://huggingface.co/codellama/CodeLlama-13b-hf">Code Llama model (13B)</a>. Please note that this model is not designed for instruction purposes but for code completion. If you're looking for instruction or want to chat with a fine-tuned model, you can use <a href="https://huggingface.co/spaces/codellama/codellama-13b-chat">this demo instead</a>. You can learn more about the model in the <a href="https://huggingface.co/blog/codellama/">blog post</a> or <a href="https://huggingface.co/papers/2308.12950">paper</a></p>
-    <p>For a chat demo of the largest Code Llama model (34B parameters), you can now <a href="https://huggingface.co/chat/">select Code Llama in Hugging Chat!</a></p>
+    <p>Compare python code generations from <a href="https://hf.co/{model_id_1}">{model_id_1}</a> (73.8% pass@1 on HumanEval) & <a href="https://hf.co/{model_id_2}">{model_id_2}</a> (73.2 pass@1 on HumanEval), which makes them surpass GPT4 (2023/03/15) on the same benchmark</p>
+    <p>Moreover, you can try those models on VSCode using HF Autocomplete extenson. Read more <a href="https://github.com/huggingface/huggingface-vscode#phind-and-wizardcoder">here</a>.</p>
+    <p>This space is cloned from <a href="https://hf.co/codellama/codellama-playground">codellama/codellama-playground</a></p>
 </div>
 """