Update description

app.py

@@ -41,18 +41,17 @@ MODEL_PATHS = {
     "type_narrative": "hallisky/lora-type-narrative-llama-3-8b",
     "type_descriptive": "hallisky/lora-type-descriptive-llama-3-8b",
 }
-FIRST_MODEL = list(MODEL_PATHS.keys())[5]
+FIRST_MODEL = list(MODEL_PATHS.keys())[0]
 MAX_NEW_TOKENS=1024
 
 DESCRIPTION = """\
-# Authorship Obfuscation
-This Space demonstrates StyleRemix, a Llama 3 model with 8B parameters fine-tuned for chat instructions. Feel free to play with it, or duplicate to run generations without a queue! If you want to run your own service, you can also [deploy the model on Inference Endpoints](https://huggingface.co/inference-endpoints).
-🔎 For more details about the Llama 2 family of models and how to use them with `transformers`, take a look [at our blog post](https://huggingface.co/blog/llama2).
-🔨 Looking for an even more powerful model? Check out the [13B version](https://huggingface.co/spaces/huggingface-projects/llama-2-13b-chat) or the large [70B model demo](https://huggingface.co/spaces/ysharma/Explore_llamav2_with_TGI).
+# Authorship Obfuscation with StyleRemix
+This Space demonstrates StyleRemix, a controllable and interpretable method for authorship obfuscation. At its core, it uses a Llama-3 model with 8B parameters and various LoRA adapters fine-tuned to rewrite text towards specific stylistic attributes (like text being longer or shorter). Feel free to play with it, or duplicate to run generations without a queue! If you want to run your own service, you can also deploy the model on [deploy the model on Inference Endpoints](https://huggingface.co/inference-endpoints). 
+<br> 🕵️ Want to learn more? Check out our paper [here](google.com) and our code [here](google.com)!
+<br> 🧐 Have questions about our work or issues with the demo? Feel free to email us at hallisky@uw.edu.
 """
 
 import subprocess
-
 def print_nvidia_smi():
     try:
         # Run the nvidia-smi command
@@ -153,6 +152,7 @@ def greet(input_text, length, function_words, grade_level, formality, sarcasm, v
         print(combo_adapter_name)
         print(list(sliders_dict.values()))
         print(list(sliders_dict.keys()))
+        print(list(model.peft_config.keys()))
 
         # Add and set the weighted adapater
         model.add_weighted_adapter(
@@ -248,7 +248,7 @@ with demo:
     gr.HTML(hide_css)
     with gr.Row():
         with gr.Column(variant="panel"):
-            gr.Markdown("# 1) Input Text\n### Enter the text to be obfuscated.")
+            gr.Markdown("# 1) Input Text\n### Enter the text to be obfuscated. We recommend *full sentences* or *paragraphs*.")
             input_text = gr.Textbox(
                 label="Input Text",
                 placeholder="The quick brown fox jumped over the lazy dogs."