MathSolve1 / app.py
Makima57's picture
Update app.py
02bf5fd verified
import gradio as gr
import ctranslate2
from transformers import AutoTokenizer
from huggingface_hub import snapshot_download
from codeexecutor import get_majority_vote
import re
# Define the model and tokenizer loading
model_prompt = "Explain and solve the following mathematical problem step by step, showing all work: "
tokenizer = AutoTokenizer.from_pretrained("AI-MO/NuminaMath-7B-TIR")
model_path = snapshot_download(repo_id="Makima57/deepseek-math-Numina")
generator = ctranslate2.Generator(model_path, device="cpu", compute_type="int8")
iterations = 10
# Function to generate predictions using the model
def get_prediction(question):
input_text = model_prompt + question
input_tokens = tokenizer.tokenize(input_text)
results = generator.generate_batch(
[input_tokens],
max_length=512,
sampling_temperature=0.7,
sampling_topk=40,
)
output_tokens = results[0].sequences[0]
predicted_answer = tokenizer.convert_tokens_to_string(output_tokens)
return predicted_answer
# Function to parse the prediction to extract the answer and steps
def parse_prediction(prediction):
lines = prediction.strip().split('\n')
answer = None
steps = []
for line in lines:
# Check for "Answer:" or "answer:"
match = re.match(r'^\s*(?:Answer|answer)\s*[:=]\s*(.*)', line)
if match:
answer = match.group(1).strip()
else:
steps.append(line)
if answer is None:
# If no "Answer:" found, assume last line is the answer
answer = lines[-1].strip()
steps = lines[:-1]
steps_text = '\n'.join(steps).strip()
return answer, steps_text
# Function to perform majority voting and get steps
def majority_vote_with_steps(question, num_iterations=10):
all_predictions = []
all_answers = []
steps_list = []
for _ in range(num_iterations):
prediction = get_prediction(question)
answer, steps = parse_prediction(prediction)
all_predictions.append(prediction)
all_answers.append(answer)
steps_list.append(steps)
# Get the majority voted answer
majority_voted_ans = get_majority_vote(all_answers)
# Find the steps corresponding to the majority voted answer
for i, ans in enumerate(all_answers):
if ans == majority_voted_ans:
steps_solution = steps_list[i]
break
else:
steps_solution = "No steps found"
return majority_voted_ans, steps_solution
# Correct solution for the problem (programmatically or hardcoded)
def fetch_correct_solution(question):
# You can define or fetch the correct solution here
# For example, for demonstration, I'll hardcode it
correct_solution_map = {
"What is 2 + 2?": "4",
"Solve for x in 2x + 3 = 7.": "x = 2",
}
return correct_solution_map.get(question, "Correct solution not available.")
# Gradio interface for user input and output
def gradio_interface(question):
final_answer, steps_solution = majority_vote_with_steps(question, iterations)
correct_solution = fetch_correct_solution(question)
return {
"Question": question,
"Majority-Voted Answer": final_answer,
"Steps to Solve": steps_solution,
"Correct Solution": correct_solution # Include the correct solution in the output
}
# Custom CSS for enhanced design (unchanged)
custom_css = """
body {
background-color: #fafafa;
font-family: 'Open Sans', sans-serif;
}
.gradio-container {
background-color: #ffffff;
border: 3px solid #007acc;
border-radius: 15px;
padding: 20px;
box-shadow: 0 8px 20px rgba(0, 0, 0, 0.15);
max-width: 800px;
margin: 50px auto;
}
h1 {
font-family: 'Poppins', sans-serif;
color: #007acc;
font-weight: bold;
font-size: 32px;
text-align: center;
margin-bottom: 20px;
}
p {
font-family: 'Roboto', sans-serif;
font-size: 18px;
color: #333;
text-align: center;
margin-bottom: 15px;
}
input, textarea {
font-family: 'Montserrat', sans-serif;
font-size: 16px;
padding: 10px;
border: 2px solid #007acc;
border-radius: 10px;
background-color: #f1f8ff;
margin-bottom: 15px;
}
#math_question {
font-size: 20px;
font-family: 'Poppins', sans-serif;
font-weight: 500px;
color: #007acc;
margin-bottom: 5px;
display: inline-block;
}
textarea {
min-height: 150px;
}
.gr-button-primary {
background-color: #007acc !important;
color: white !important;
border-radius: 10px !important;
font-size: 18px !important;
font-weight: bold !important;
padding: 10px 20px !important;
font-family: 'Montserrat', sans-serif !important;
transition: background-color 0.3s ease !important;
}
.gr-button-primary:hover {
background-color: #005f99 !important;
}
.gr-button-secondary {
background-color: #f44336 !important;
color: white !important;
border-radius: 10px !important;
font-size: 18px !important;
font-weight: bold !important;
padding: 10px 20px !important;
font-family: 'Montserrat', sans-serif !important;
transition: background-color 0.3s ease !important;
}
.gr-button-secondary:hover {
background-color: #c62828 !important;
}
.gr-output {
background-color: #e0f7fa;
border: 2px solid #007acc;
border-radius: 10px;
padding: 15px;
font-size: 16px;
font-family: 'Roboto', sans-serif;
font-weight: bold;
color: #00796b;
}
"""
# Gradio app setup
interface = gr.Interface(
fn=gradio_interface,
inputs=[
gr.Textbox(label="🧠 Math Question", placeholder="Enter your math question here...", elem_id="math_question"),
],
outputs=[
gr.JSON(label="📊 Results"), # Display the results in a JSON format
],
title="🔢 Math Question Solver",
description="Enter a math question to get the model's majority-voted answer, steps to solve the problem, and the correct solution.",
css=custom_css # Apply custom CSS
)
if __name__ == "__main__":
interface.launch()