Initial commit

Tensorflow/tutoriel33/graph.py

@@ -0,0 +1,68 @@
+import matplotlib.pyplot as plt
+import matplotlib.ticker as mtick
+import csv
+import sys
+import numpy as np
+
+if len(sys.argv)!=2:
+    print("Usage:", sys.argv[0], "<fichier csv>")
+    quit()
+fichier=sys.argv[1]
+
+def calc(tab_data, fenetre):
+    tab_m=[]
+    for i in range(len(tab_data)-fenetre):
+        m=np.mean(tab_data[i:i+fenetre])
+        tab_m.append(m)
+    return tab_m
+
+x=[]
+accuracy=[]
+loss=[]
+val_accuracy=[]
+val_loss=[]
+fenetre=50
+
+val=0
+with open(fichier,'r') as csvfile:
+    plots=csv.reader(csvfile, delimiter=',')
+    next(plots)
+    for row in plots:
+        x.append(float(row[0]))
+        accuracy.append(float(row[1]))
+        loss.append(float(row[2]))
+        if len(row)==5:
+            val_accuracy.append(float(row[3]))
+            val_loss.append(float(row[4]))
+            val=1
+
+
+fig, (ax1, ax2)=plt.subplots(2)
+fig.set_size_inches(9, 7, forward=True)
+
+ax1.set_ylim([0, 1.0])
+ax1.grid(which='both')
+ax1.yaxis.set_major_formatter(mtick.PercentFormatter(1.0))
+ln=ax1.plot(x, accuracy, label='Accuracy')
+
+ax1_=ax1.twinx()
+ax1_.set_ylim([0.0, 2.0])
+ln_=ax1_.plot(x, loss, label='Loss', color='red')
+
+lns=ln+ln_
+labs=[l.get_label() for l in lns]
+
+ax2.set_ylim([0, 1.0])
+ax2.grid(which='both')
+ax2.yaxis.set_major_formatter(mtick.PercentFormatter(1.0))
+ln=ax2.plot(x, val_accuracy, label='Val accuracy')
+
+ax2_=ax2.twinx()
+ax2_.set_ylim([0.0, 2.0])
+ln_=ax2_.plot(x, val_loss, label='Val loss', color='red')
+
+lns=ln+ln_
+labs=[l.get_label() for l in lns]
+ax2.legend(lns, labs, loc='upper center', bbox_to_anchor=(0.5, -0.1), fancybox=True, shadow=True, ncol=5)
+
+plt.show()