GPT-3 Code modification: plot w1_vals, w2_vals using plt

SGD.py

#write the code for an efficient and modular, reentrant, Stochastic gradient descent in python for any set of two unknown float inputs. output the results in matlab graph format and as an Open Office Spreadsheet.

import numpy as np
import matplotlib.pyplot as plt
%matplotlib inline

def sgd(w1, w2, eta, n_iter):
    """
    Performs stochastic gradient descent (SGD) using two unknown float inputs.
    
    Parameters
    ----------
    w1 : float
        First weight.
    w2 : float
        Second weight.
    eta : float
        Learning rate.
    n_iter : int
        Number of iterations.
    
    Returns
    -------
    w1_vals : list
        Values of w1 across iterations.
    w2_vals : list
        Values of w2 across iterations.
    """
    # Initialize lists to store weights
    w1_vals = []
    w2_vals = []
    
    # Iterate for n_iter
    for i in range(n_iter):
        # Calculate current prediction
        pred = w1 * x1 + w2 * x2
        
        # Calculate current error
        error = pred - y
        
        # Update weights
        w1 = w1 - eta * error * x1
        w2 = w2 - eta * error * x2
        
        # Store weights
        w1_vals.append(w1)
        w2_vals.append(w2)
    
    return w1_vals, w2_vals

# Instantiate inputs x1 and x2
x1 = 0.5
x2 = 0.5

# Instantiate output y
y = 1.0

# Set learning rate and number of iterations
eta = 0.05
n_iter = 1000

# Calculate w1_vals and w2_vals using sgd
w1_vals, w2_vals = sgd(x1, x2, eta, n_iter)

# Extract each element of w1_vals as a separate list
w1_vals = [val[0] for val in w1_vals]

# Extract each element of w2_vals as a separate list
w2_vals = [val[1] for val in w2_vals]

# Create shapefile of closed curves from the list of lists of point x and y coordinates
plt.plot(w1_vals, w2_vals)

# Show plot
plt.show()

#   END CODE
 
end = time()
print ("------------------------", (end-start)/60, " minutes-----------------------------", (end-start)/60/60, " hours")