Update app.py

app.py

@@ -1,59 +1,4 @@
-import os
-import json
-import subprocess
-from threading import Thread
-
-import torch
-import spaces
-import gradio as gr
-from transformers import AutoModelForCausalLM, AutoTokenizer, BitsAndBytesConfig, TextIteratorStreamer
-
-subprocess.run('pip install flash-attn --no-build-isolation', env={'FLASH_ATTENTION_SKIP_CUDA_BUILD': "TRUE"}, shell=True)
-
-MODEL_ID = "nikravan/Marco-O1-q4"
-CHAT_TEMPLATE = "ChatML"
-MODEL_NAME = MODEL_ID.split("/")[-1]
-CONTEXT_LENGTH = 16000
-
-# Estableciendo valores directamente para las variables
-COLOR = "blue"  # Color predeterminado de la interfaz
-EMOJI = "🤖"  # Emoji predeterminado para el modelo
-DESCRIPTION = f"This is the {MODEL_NAME} model designed for testing thinking for general AI tasks."  # Descripción predeterminada
-
-latex_delimiters_set = [{
-        "left": "\\(",
-        "right": "\\)",
-        "display": False 
-    }, {
-        "left": "\\begin{equation}",
-        "right": "\\end{equation}",
-        "display": True 
-    }, {
-        "left": "\\begin{align}",
-        "right": "\\end{align}",
-        "display": True
-    }, {
-        "left": "\\begin{alignat}",
-        "right": "\\end{alignat}",
-        "display": True
-    }, {
-        "left": "\\begin{gather}",
-        "right": "\\end{gather}",
-        "display": True
-    }, {
-        "left": "\\begin{CD}",
-        "right": "\\end{CD}",
-        "display": True
-    }, {
-        "left": "\\[",
-        "right": "\\]",
-        "display": True
-    }]
-
-
-@spaces.GPU()
 def predict(message, history, system_prompt, temperature, max_new_tokens, top_k, repetition_penalty, top_p):
-    # Format history with a given chat template
     if CHAT_TEMPLATE == "Auto":
         stop_tokens = [tokenizer.eos_token_id]
         instruction = system_prompt + "\n\n"
@@ -102,31 +47,16 @@ def predict(message, history, system_prompt, temperature, max_new_tokens, top_k,
         outputs.append(new_token)
         if new_token in stop_tokens:
             break
-        yield "".join(outputs)
-
+        result = "".join(outputs)
+        # Wrapping result in Markdown for LaTeX rendering
+        yield f"$$ {result} $$"
 
-# Load model
-device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
-quantization_config = BitsAndBytesConfig(
-    load_in_4bit=True,
-    bnb_4bit_compute_dtype=torch.bfloat16
-)
-tokenizer = AutoTokenizer.from_pretrained('AIDC-AI/Marco-o1')
-model = AutoModelForCausalLM.from_pretrained(
-    MODEL_ID,
-    device_map="auto",
-    quantization_config=quantization_config,
-    attn_implementation="flash_attention_2",
-)
 
 # Create Gradio interface
 gr.ChatInterface(
     predict,
     title=EMOJI + " " + MODEL_NAME,
     description=DESCRIPTION,
-    
-
-     
     additional_inputs_accordion=gr.Accordion(label="⚙️ Parameters", open=False),
     additional_inputs=[
         gr.Textbox("You are a code assistant.", label="System prompt"),
@@ -138,4 +68,3 @@ gr.ChatInterface(
     ],
     theme=gr.themes.Soft(primary_hue=COLOR),
 ).queue().launch()
-