Update app.py

app.py

@@ -33,7 +33,7 @@ sft_start_token =  "<|im_start|>"
 sft_end_token = "<|im_end|>"
 ct_end_token = "<|endoftext|>"
 
-system_prompt= 'You are a CodeLLM developed by INF.'
+# system_prompt= 'You are a CodeLLM developed by INF.'
 
 
 # Function to generate model predictions.
@@ -45,8 +45,8 @@ def predict(message, history):
         stop = StopOnTokens()
     
         model_messages = []
-        print(f'history: {history}')
-        model_messages.append({"role": system_role, "content": system_prompt})
+        # print(f'history: {history}')
+
         for i, item in enumerate(history):
             model_messages.append({"role": user_role, "content": item[0]})
             model_messages.append({"role": assistant_role, "content": item[1]})
@@ -55,7 +55,7 @@ def predict(message, history):
         
         print(f'model_messages: {model_messages}')
     
-        print(f'model_final_inputs: {tokenizer.apply_chat_template(model_messages, add_generation_prompt=True, tokenize=False)}', flush=True)
+        # print(f'model_final_inputs: {tokenizer.apply_chat_template(model_messages, add_generation_prompt=True, tokenize=False)}', flush=True)
         model_inputs = tokenizer.apply_chat_template(model_messages, add_generation_prompt=True, return_tensors="pt").to(device)
         # model_inputs = tokenizer([messages], return_tensors="pt").to(device)
         
@@ -64,7 +64,8 @@ def predict(message, history):
             input_ids=model_inputs,
             streamer=streamer,
             max_new_tokens=1024,
-            do_sample=False
+            do_sample=False,
+            stopping_criteria=StoppingCriteriaList([stop])
         )
         
         t = Thread(target=model.generate, kwargs=generate_kwargs)
@@ -87,24 +88,21 @@ full-height {
 """
 
 prompt_examples = [
-    'How to cook a fish?',
-    'Cara memanggang ikan',
-    'วิธีย่างปลา',
-    'Cách nướng cá'
+    'Write a quick sort algorithm in python.',
+    'Write a colorful greedy snake game using pygame.',
+    'How to use numpy?'
 ]
 
 placeholder = """
 <div style="opacity: 0.5;">
-    <img src="https://raw.githubusercontent.com/sail-sg/sailor-llm/main/misc/banner.jpg" style="width:30%;">
-    <br>Sailor models are designed to understand and generate text across diverse linguistic landscapes of these SEA regions:
-    <br>🇮🇩Indonesian, 🇹🇭Thai, 🇻🇳Vietnamese, 🇲🇾Malay, and 🇱🇦Lao.
+    <img src="https://github.com/OpenCoder-llm/opencoder-llm.github.io/blob/main/static/images/opencoder_icon.jpg?raw=true" style="width:30%;">
 </div>
 """
 
-chatbot = gr.Chatbot(label='OpenCoder', placeholder=None) 
+
+chatbot = gr.Chatbot(label='OpenCoder', placeholder=placeholder) 
 with gr.Blocks(theme=gr.themes.Soft(), fill_height=True) as demo:
-    # gr.Markdown("""<center><font size=8>Sailor-Chat Bot⚓</center>""")
-    gr.Markdown("""<p align="center"><img src="https://github.com/sail-sg/sailor-llm/raw/main/misc/wide_sailor_banner.jpg" style="height: 110px"/><p>""")
+    
     gr.ChatInterface(predict, chatbot=chatbot, fill_height=True, examples=prompt_examples, css=css)
 
     demo.launch()  # Launching the web interface.