Update app.py

app.py

@@ -1,38 +1,33 @@
 import gradio as gr
 import os
 import spaces
-from transformers import GemmaTokenizer, AutoModelForCausalLM
-from transformers import AutoModelForCausalLM, AutoTokenizer, TextIteratorStreamer
+from transformers import AutoTokenizer, AutoModelForCausalLM, TextIteratorStreamer
 from threading import Thread
 
 # Set an environment variable
-# HF_TOKEN = os.environ.get("HF_TOKEN", None)
-
+HF_TOKEN = os.environ.get("HF_TOKEN", None)
 
 DESCRIPTION = '''
 <div>
-<h1 style="text-align: center;">Meta Llama3 8B</h1>
-<p>This Space demonstrates the instruction-tuned model <a href="https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct"><b>Meta Llama3 8b Chat</b></a>. Meta Llama3 is the new open LLM and comes in two sizes: 8b and 70b. Feel free to play with it, or duplicate to run privately!</p>
-<p>🔎 For more details about the Llama3 release and how to use the model with <code>transformers</code>, take a look <a href="https://huggingface.co/blog/llama3">at our blog post</a>.</p>
-<p>🦕 Looking for an even more powerful model? Check out the <a href="https://huggingface.co/chat/"><b>Hugging Chat</b></a> integration for Meta Llama 3 70b</p>
+<h1 style="text-align: center;">StevenChen16 Llama3 8B Lawyer</h1>
+<p>This Space demonstrates the instruction-tuned model <a href="https://huggingface.co/StevenChen16/llama3-8b-Lawyer"><b>StevenChen16 Llama3 8B Lawyer</b></a>. This model is fine-tuned to provide legal advice based on US and Canada law.</p>
+<p>Feel free to play with it, or duplicate to run privately!</p>
 </div>
 '''
 
 LICENSE = """
 <p/>
 ---
-Built with Meta Llama 3
+Built with StevenChen16 Llama3 8B Lawyer
 """
 
 PLACEHOLDER = """
 <div style="padding: 30px; text-align: center; display: flex; flex-direction: column; align-items: center;">
-   <img src="https://ysharma-dummy-chat-app.hf.space/file=/tmp/gradio/8e75e61cc9bab22b7ce3dec85ab0e6db1da5d107/Meta_lockup_positive%20primary_RGB.jpg" style="width: 80%; max-width: 550px; height: auto; opacity: 0.55;  "> 
-   <h1 style="font-size: 28px; margin-bottom: 2px; opacity: 0.55;">Meta llama3</h1>
-   <p style="font-size: 18px; margin-bottom: 2px; opacity: 0.65;">Ask me anything...</p>
+   <h1 style="font-size: 28px; margin-bottom: 2px; opacity: 0.55;">StevenChen16 Llama3 8B Lawyer</h1>
+   <p style="font-size: 18px; margin-bottom: 2px; opacity: 0.65;">Ask me anything about US and Canada law...</p>
 </div>
 """
 
-
 css = """
 h1 {
   text-align: center;
@@ -48,10 +43,11 @@ h1 {
 
 # Load the tokenizer and model
 tokenizer = AutoTokenizer.from_pretrained("StevenChen16/llama3-8b-Lawyer")
-model = AutoModelForCausalLM.from_pretrained("StevenChen16/llama3-8b-Lawyer", device_map="auto")  # to("cuda:0") 
+model = AutoModelForCausalLM.from_pretrained("StevenChen16/llama3-8b-Lawyer", device_map="auto")
+
 terminators = [
     tokenizer.eos_token_id,
-    tokenizer.convert_tokens_to_ids("<|eot_id|>")
+    tokenizer.convert_tokens_to_ids("")
 ]
 
 @spaces.GPU(duration=120)
@@ -97,12 +93,11 @@ def chat_llama3_8b(message: str,
     outputs = []
     for text in streamer:
         outputs.append(text)
-        #print(outputs)
         yield "".join(outputs)
         
 
 # Gradio block
-chatbot=gr.Chatbot(height=450, placeholder=PLACEHOLDER, label='Gradio ChatInterface')
+chatbot = gr.Chatbot(height=450, placeholder=PLACEHOLDER, label='Gradio ChatInterface')
 
 with gr.Blocks(fill_height=True, css=css) as demo:
     
@@ -140,4 +135,4 @@ with gr.Blocks(fill_height=True, css=css) as demo:
     gr.Markdown(LICENSE)
     
 if __name__ == "__main__":
-    demo.launch()
+    demo.launch()