Helpers.py

from IPython.display import clear_output
from matplotlib import pyplot as plt
import collections
import numpy as np
import pandas as pd
from PIL import Image
import torch
import torchvision
from sklearn.metrics import accuracy_score

def live_plot(loss, train_acc, valid_acc=None, figsize=(7,5), title=''):
    clear_output(wait=True)
    fig, ax1 = plt.subplots(figsize=figsize)
    ax1.plot(loss, label='Training Loss', color='red')
    ax1.legend(loc='lower left')
    ax1.set_ylabel('Cross Entropy Loss')
    ax2 = ax1.twinx()
    ax2.plot(train_acc, label='Training Accuracy', color='green')
    if valid_acc is not None:
        ax2.plot(valid_acc, label='Validation Accuracy', color='blue')
    ax2.legend(loc='lower right')
    ax2.set_ylabel('Accuracy (%)')
    ax2.set_xlabel('Epoch')
    plt.title(title)
    plt.show()