Update app.py

app.py

@@ -5,47 +5,51 @@ from huggingface_hub import hf_hub_download
 
 model = Llama(
     model_path=hf_hub_download(
-        repo_id=os.environ.get("REPO_ID", "bartowski/QwQ-32B-Preview-GGUF"),
-        filename=os.environ.get("MODEL_FILE", "QwQ-32B-Preview-Q3_K_L.gguf"),
+        repo_id=os.environ.get("REPO_ID", "Lyte/LLaMA-O1-Supervised-1129-Q4_K_M-GGUF"),
+        filename=os.environ.get("MODEL_FILE", "llama-o1-supervised-1129-q4_k_m.gguf"),
     )
 )
 
 DESCRIPTION = '''
-# QwQ-32B-Preview | Duplicate the space and set it to private for faster & personal inference for free.
-Qwen/QwQ-32B-Preview: an experimental research model developed by the Qwen Team.  
+# SimpleBerry/LLaMA-O1-Supervised-1129 | Duplicate the space and set it to private for faster & personal inference for free.
+SimpleBerry/LLaMA-O1-Supervised-1129: an experimental research model developed by the SimpleBerry.  
 Focused on advancing AI reasoning capabilities.  
 
 **To start a new chat**, click "clear" and start a new dialog.
 '''
 
 LICENSE = """
---- Apache 2.0 License ---
+--- MIT License ---
 """
 
+template = "<start_of_father_id>-1<end_of_father_id><start_of_local_id>0<end_of_local_id><start_of_thought><problem>{content}<end_of_thought><start_of_rating><positive_rating><end_of_rating>\n<start_of_father_id>0<end_of_father_id><start_of_local_id>1<end_of_local_id><start_of_thought><expansion>"
+
+def llama_o1_template(data):
+    #query = data['query']
+    text = template.format(content=data)
+    return text
+
 def generate_text(message, history, max_tokens=512, temperature=0.9, top_p=0.95):
-    """Generate a response using the Llama model."""
     temp = ""
-    response = model.create_chat_completion(
-        messages=[{"role": "system", "content": "You are a helpful and harmless assistant. You are Qwen developed by Alibaba. You should think step-by-step."},
-                  {"role": "user", "content": message}],
-        temperature=temperature,
-        max_tokens=max_tokens,
-        top_p=top_p,
-        stream=True,
-    )
-    for streamed in response:
-        delta = streamed["choices"][0].get("delta", {})
-        text_chunk = delta.get("content", "")
-        temp += text_chunk
+    input_texts = [llama_o1_template(message)]
+    input_texts = [input_text.replace('<|end_of_text|>','') for input_text in input_texts]
+    print(f"input_texts[0]: {input_texts[0]}")
+    inputs = model.tokenize(input_texts[0].encode('utf-8'))
+    for token in model.generate(inputs, top_p=top_p, temp=temperature):
+        print(f"token: {token}")
+        text = model.detokenize([token])
+        print(f"text detok: {text}")
+        temp += text.decode('utf-8')
         yield temp
 
+
 with gr.Blocks() as demo:
     gr.Markdown(DESCRIPTION)
 
     chatbot = gr.ChatInterface(
         generate_text,
-        title="Qwen/QwQ-32B-Preview | GGUF Demo",
-        description=" settings below if needed.",
+        title="SimpleBerry/LLaMA-O1-Supervised-1129 | GGUF Demo",
+        description="Edit Settings below if needed.",
         examples=[
             ["How many r's are in the word strawberry?"],
             ['What is the most optimal way to do Test-Time Scaling?'],
@@ -56,9 +60,9 @@ with gr.Blocks() as demo:
     )
 
     with gr.Accordion("Adjust Parameters", open=False):
-        gr.Slider(minimum=512, maximum=4096, value=1024, step=1, label="Max Tokens")
-        gr.Slider(minimum=0.1, maximum=1.5, value=0.9, step=0.1, label="Temperature")
-        gr.Slider(minimum=0.1, maximum=1.0, value=0.95, step=0.05, label="Top-p (nucleus sampling)")
+        gr.Slider(minimum=1024, maximum=8192, value=2048, step=1, label="Max Tokens")
+        gr.Slider(minimum=0.1, maximum=1.5, value=0.7, step=0.1, label="Temperature")
+        gr.Slider(minimum=0.05, maximum=1.0, value=0.95, step=0.01, label="Top-p (nucleus sampling)")
 
     gr.Markdown(LICENSE)