import sys
import argparse
import numpy as np
import matplotlib.pyplot as plt
import matplotlib.gridspec as gridspec
def plot_charts(modelLoc):
fileObj = open("./" + modelLoc + "/log.txt", 'r')
epoch = []
trainLoss = []
trainScore = []
valLoss = []
valScore = []
# Read in File
for line in fileObj:
words = line.split()
if words[0] == 'epoch':
epoch.append(int(words[1][:-1]))
elif words[0] == 'train_loss:':
trainLoss.append(float(words[1][:-1]))
trainScore.append(float(words[3]))
elif words[0] == 'eval':
valLoss.append(float(words[2][:-1]))
valScore.append(float(words[4]))
fileObj.close()
minValLoss = min(valLoss)
epochValLoss = np.argmin(valLoss)
maxValScore = max(valScore)
epochValScore = np.argmax(valScore)
# Plot Loss and Score
plt.figure()
plt.suptitle(modelLoc)
# train/val loss
plt.subplot(211)
plt.plot(epoch, trainLoss, label='Training')
plt.plot(epoch, valLoss, label='Validation')
plt.plot(epochValLoss, minValLoss, marker='x', markersize=3, color="black")
plt.xlabel('loss')
plt.ylabel('score')
plt.title('Training and Validation Loss')
plt.legend()
plt.minorticks_on()
plt.grid(True, which='both')
# train/val score
plt.subplot(212)
plt.plot(epoch, trainScore, label='Training')
plt.plot(epoch, valScore, label='Validation')
plt.plot(epochValScore, maxValScore, marker='x', markersize=3, color="black")
plt.xlabel('epochs')
plt.ylabel('score')
plt.title('Training and Validation Score')
plt.legend()
plt.minorticks_on()
plt.grid(True, which='both')
plt.subplots_adjust(hspace=0.5)
#plt.show()
plt.savefig("./" + modelLoc + ".png")
if __name__ == '__main__':
modelLoc = sys.argv[1]
plot_charts(modelLoc)
