Update app.py

app.py

@@ -24,7 +24,6 @@ PLACEHOLDER = """
 </center>
 """
 
-
 CSS = """
 .duplicate-button {
     margin: auto !important;
@@ -37,9 +36,11 @@ h3 {
 }
 .message-wrap {
     overflow-x: auto;
+    white-space: pre-wrap !important;
 }
 .message-wrap p {
     margin-bottom: 1em;
+    white-space: pre-wrap !important;
 }
 .message-wrap pre {
     background-color: #f6f8fa;
@@ -54,13 +55,14 @@ h3 {
     font-family: monospace;
 }
 """
-device = "cuda" # for GPU usage or "cpu" for CPU usage
+
+device = "cuda"  # for GPU usage or "cpu" for CPU usage
 
 quantization_config = BitsAndBytesConfig(
     load_in_4bit=True,
     bnb_4bit_compute_dtype=torch.bfloat16,
     bnb_4bit_use_double_quant=True,
-    bnb_4bit_quant_type= "nf4")
+    bnb_4bit_quant_type="nf4")
 
 tokenizer = AutoTokenizer.from_pretrained(MODEL)
 model = AutoModelForCausalLM.from_pretrained(
@@ -74,6 +76,14 @@ model = AutoModelForCausalLM.from_pretrained(
 if tokenizer.pad_token_id is None:
     tokenizer.pad_token_id = tokenizer.eos_token_id
 
+def format_text(text):
+    """Helper function to format text with proper line breaks and spacing"""
+    # Replace single newlines with double newlines for paragraph spacing
+    formatted = text.replace('\n', '\n\n')
+    # Remove extra spaces between paragraphs
+    formatted = '\n'.join(line.strip() for line in formatted.split('\n'))
+    return formatted
+
 @spaces.GPU()
 def stream_chat(
     message: str, 
@@ -99,85 +109,104 @@ def stream_chat(
 
     conversation.append({"role": "user", "content": message})
 
-    input_ids = tokenizer.apply_chat_template(conversation, add_generation_prompt=True, return_tensors="pt").to(model.device)
+    input_ids = tokenizer.apply_chat_template(
+        conversation, 
+        add_generation_prompt=True, 
+        return_tensors="pt"
+    ).to(model.device)
     
-    streamer = TextIteratorStreamer(tokenizer, timeout=60.0, skip_prompt=True, skip_special_tokens=True)
+    streamer = TextIteratorStreamer(
+        tokenizer, 
+        timeout=60.0, 
+        skip_prompt=True, 
+        skip_special_tokens=True
+    )
     
     generate_kwargs = dict(
         input_ids=input_ids, 
-        max_new_tokens = max_new_tokens,
-        do_sample = False if temperature == 0 else True,
-        top_p = top_p,
-        top_k = top_k,
-        eos_token_id = tokenizer.eos_token_id,
-        pad_token_id = tokenizer.pad_token_id,
-        temperature = temperature,
+        max_new_tokens=max_new_tokens,
+        do_sample=False if temperature == 0 else True,
+        top_p=top_p,
+        top_k=top_k,
+        eos_token_id=tokenizer.eos_token_id,
+        pad_token_id=tokenizer.pad_token_id,
+        temperature=temperature,
         repetition_penalty=penalty,
         streamer=streamer,
     )
 
+    buffer = ""
+    current_line = ""
+    
     with torch.no_grad():
         thread = Thread(target=model.generate, kwargs=generate_kwargs)
         thread.start()
         
-    buffer = ""
     for new_text in streamer:
+        # Add the new text to both buffers
         buffer += new_text
-        yield buffer
+        current_line += new_text
+        
+        # Check if we have complete lines to process
+        if '\n' in current_line:
+            lines = current_line.split('\n')
+            # The last element might be incomplete, so keep it in current_line
+            current_line = lines[-1]
+            # Format the complete text
+            formatted_buffer = format_text(buffer)
+            yield formatted_buffer
+        else:
+            yield buffer
 
             
-chatbot = gr.Chatbot(height=600, placeholder=PLACEHOLDER)
-
-with gr.Blocks(css=CSS, theme="soft") as demo:
-    gr.HTML(TITLE)
-    gr.DuplicateButton(value="Duplicate Space for private use", elem_classes="duplicate-button")
-    gr.ChatInterface(
-        fn=stream_chat,
-        chatbot=chatbot,
-        fill_height=True,
-        additional_inputs_accordion=gr.Accordion(label="⚙️ Parameters", open=False, render=False),
-        additional_inputs=[
-            gr.Textbox(
-                value="""You are an AI expert at providing high-quality answers. Your process involves these steps:
+chatbot = gr.Chatbot(
+    height=600,
+    placeholder=PLACEHOLDER,
+    bubble_full_width=False,
+    show_copy_button=True
+)
 
+DEFAULT_SYSTEM_PROMPT = """You are an AI expert at providing high-quality answers. Your process involves these steps:
 1. Initial Thought: Use the <Thinking> tag to reason step-by-step and generate your best possible response to the following request: [User's Request Here].
 Example:
 <Thinking> Step 1: Understand the request. Step 2: Analyze potential solutions. Step 3: Choose the optimal response. </Thinking>
-
-
 2. Self-Critique: Critically evaluate your initial response within <Critique> tags, focusing on:
-
 Accuracy: Is it factually correct and verifiable?
-
 Clarity: Is it easy to understand and free of ambiguity?
-
 Completeness: Does it fully address the user's request?
-
 Improvement: What specific aspects could be better?
 Example:
 <Critique> Accuracy: Verified. Clarity: Needs simplification. Completeness: Add examples. </Critique>
-
-
-
 3. Revision: Based on your critique, use <Revising> tags to refine and improve your response.
 Example:
 <Revising> Adjusting for clarity and adding an example to improve understanding. </Revising>
-
-
 4. Final Response: Present your revised answer clearly within <Final> tags.
 Example:
 <Final> This is the improved response. </Final>
-
-
 5. Tag Innovation: If necessary, create and define new tags to better structure your reasoning or enhance clarity. Use them consistently.
 Example:
 <Definition> This tag defines a new term introduced in the response. </Definition>
+Ensure every part of your thought process and output is properly enclosed in appropriate tags for clarity and organization."""
 
-
-
-Ensure every part of your thought process and output is properly enclosed in appropriate tags for clarity and organization.
-
-""",
+with gr.Blocks(css=CSS, theme="soft") as demo:
+    gr.HTML(TITLE)
+    gr.DuplicateButton(
+        value="Duplicate Space for private use",
+        elem_classes="duplicate-button"
+    )
+    
+    gr.ChatInterface(
+        fn=stream_chat,
+        chatbot=chatbot,
+        fill_height=True,
+        additional_inputs_accordion=gr.Accordion(
+            label="⚙️ Parameters",
+            open=False,
+            render=False
+        ),
+        additional_inputs=[
+            gr.Textbox(
+                value=DEFAULT_SYSTEM_PROMPT,
                 label="System Prompt",
                 lines=5,
                 render=False,
@@ -194,7 +223,7 @@ Ensure every part of your thought process and output is properly enclosed in app
                 minimum=128,
                 maximum=32000,
                 step=1,
-                value= 8192,
+                value=8192,
                 label="Max new tokens",
                 render=False,
             ),
@@ -224,14 +253,13 @@ Ensure every part of your thought process and output is properly enclosed in app
             ),
         ],
         examples=[
-            ["What is meant by a Singularity? "],
-            ["Explain the theory of Relativty"],
-            ["Explain your thought process"],
-            ["Explain how mamba2 structure LLMs work and how do they differ from transformers? "],
+            ["What is meant by a Singularity?"],
+            ["Explain the theory of Relativity"],
+            ["Explain your thought process in details"],
+            ["Explain how mamba2 structure LLMs work and how do they differ from transformers?"],
         ],
         cache_examples=False,
     )
 
-
 if __name__ == "__main__":
     demo.launch()