README.md

metadata

dataset_info:
  features:
    - name: instruction
      dtype: string
    - name: output
      dtype: string
  splits:
    - name: train
      num_bytes: 15268888.05
      num_examples: 487500
    - name: test
      num_bytes: 391509.95
      num_examples: 12500
  download_size: 12160789
  dataset_size: 15660398
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: test
        path: data/test-*

Simple Math

Just like my teacher gave me homework, i thought maybe we can also add some of these basics on the trainings of our models.

It was created with this code, if you add more complex operations and so.. please share the code :D thank you

import random
# Define the number of samples you want to generate
num_samples = 500000
# Define the range for the random numbers
min_value = -99.99
max_value = 99.99
# Define the arithmetic operations
operations = ['+', '-', '*', '/']
# Generate data
data = []
for _ in range(num_samples):
    num1 = float("%.3f" % random.uniform(min_value, max_value))
    num2 = float("%.3f" % random.uniform(min_value, max_value))
    while num2 == 0.0:
         num2 = float("%.3f" % random.uniform(min_value, max_value))
    while num1 == 0.0:
         num1 = float("%.3f" % random.uniform(min_value, max_value))
    operation = random.choice(operations)
    if operation == '/':
        result = num1 / num2
    elif operation == '-':
        result = num1 - num2
    elif operation == '*':
        result = num1 * num2
    elif operation == '+':
        result = num1 + num2
    output = "%.4f" % result
    instruction = f"{num1} {operation} {num2}"
    data.append({'instruction': instruction, 'output': output})
# Create the dataset
import json
out_file = 'arithmetic-float4a.json'
with open(out_file, 'w') as f:
    json.dump(data, f)

We have also a new community submission for a more complex dataset of 3 elements:

import random
import json
import operator

# Define the number of samples you want to generate
num_samples = 500000
# Define the range for the random numbers
min_value = -99.99
max_value = 99.99
# Define the arithmetic operations and their corresponding functions
operations = {
    '+': operator.add,
    '-': operator.sub,
    '*': operator.mul,
    '/': operator.truediv
}

def generate_random_number():
    return float("%.3f" % random.uniform(min_value, max_value))

def safe_division(numerator, denominator):
    return numerator if denominator == 0 else numerator / denominator

# Generate complex data
data = []
for _ in range(num_samples):
    num1 = generate_random_number()
    num2 = generate_random_number()
    num3 = generate_random_number()
    # Ensure num2 and num3 are not zero if they will be used as divisors
    if num2 == 0.0:
        num2 = generate_random_number()
    if num3 == 0.0:
        num3 = generate_random_number()

    # Randomly choose two operations
    operation1 = random.choice(list(operations.keys()))
    operation2 = random.choice(list(operations.keys()))
    
    # Create a more complex expression
    if random.choice([True, False]):
        # With parentheses
        expression = f"({num1} {operation1} {num2}) {operation2} {num3}"
        if operation1 == '/':
            intermediate_result = safe_division(num1, num2)
        else:
            intermediate_result = operations[operation1](num1, num2)
        if operation2 == '/' and intermediate_result != 0:
            result = safe_division(intermediate_result, num3)
        else:
            result = operations[operation2](intermediate_result, num3)
    else:
        # Without parentheses
        expression = f"{num1} {operation1} {num2} {operation2} {num3}"
        if operation1 == '/':
            intermediate_result = safe_division(num1, num2)
        else:
            intermediate_result = operations[operation1](num1, num2)
        if operation2 == '/':
            result = safe_division(intermediate_result, num3)
        else:
            result = operations[operation2](intermediate_result, num3)

    output = "%.4f" % result
    data.append({'instruction': expression, 'output': output})

# Create the dataset
out_file = 'arithmetic-float-complex.json'
with open(out_file, 'w') as f:
    json.dump(data, f)

If you use Simple Math o train your model, please cite on the modelcard or the paper. Thank you