loss_plot.py

import re
import matplotlib.pyplot as plt
import pandas as pd

loss_filenames = ["losses_sim_1575781367.650115.txt", "losses_vgg_1575624281.130714.txt", "losses_1575458071.3440142.txt"]

#loss_filename = "losses_1575458071.3440142.txt"
#loss_filename = "losses_vgg_1575624281.130714.txt"
#loss_filename = "losses_sim_1575781367.650115.txt"

for loss_filename in loss_filenames:

  rgx = r'\[[-?0-9.]*\]'

  disc_loss = []
  gen_loss = []

  with open(loss_filename, "r+") as f:
    for line in f.readlines():
      try:
        nums = re.findall(rgx, line)
        disc = float(nums[2][1:-1])
        gen = float(nums[3][1:-1])
        disc_loss.append(disc)
        gen_loss.append(gen)
      except IndexError:
        pass
  
  #df=pd.DataFrame({'x_disc': range(len(disc_loss)), 'x_gen': range(len(gen_loss)), 'disc_loss': disc_loss, 'gen_loss': gen_loss})

  #plt.plot('x_disc', 'disc_loss', data=df, marker='', color='blue', linewidth=2)
  #plt.plot('x_gen', 'gen_loss', data=df, marker='', color='orange', linewidth=2)
  #plt.legend()
  plt.plot(range(len(disc_loss)),disc_loss)
  plt.xlabel('iterations')
  plt.ylabel('loss')
  plt.title('Discriminator Model Loss')
  plt.legend(['disc loss'])
  plt.savefig(loss_filename[:-4]+'_discriminator'+'.png')
  print('Saved figure at {}'.format(loss_filename[:-4]+'_discriminator'+'.png'))
  plt.close()

  plt.plot(range(len(gen_loss)), gen_loss)
  plt.xlabel('iterations')
  plt.ylabel('loss')
  if loss_filename[7] == 's':
    plt.ylim(-14, 1)
  plt.title('Generator Model Loss')
  plt.legend(['gen loss'])
  plt.savefig(loss_filename[:-4]+'_generator'+'.png')
  print('Saved figure at {}'.format(loss_filename[:-4]+'_generator'+'.png'))
  plt.close()