changement des base d'entrainements

code/fonctions_activations_classiques/sin.py

 
 model_sin.add(tf.keras.Input(shape=(1,)))
 
-model_sin.add(tf.keras.layers.Dense(64, activation=sin))
-model_sin.add(tf.keras.layers.Dense(64, activation=sin))
-model_sin.add(tf.keras.layers.Dense(64, activation=sin))
-model_sin.add(tf.keras.layers.Dense(64, activation=sin))
-model_sin.add(tf.keras.layers.Dense(64, activation=sin))
-model_sin.add(tf.keras.layers.Dense(64, activation=sin))
-# model_sin.add(tf.keras.layers.Dense(8, activation=sin))
+# model_sin.add(tf.keras.layers.Dense(64, activation=sin))
+# model_sin.add(tf.keras.layers.Dense(64, activation=sin))
+# model_sin.add(tf.keras.layers.Dense(64, activation=sin))
+# model_sin.add(tf.keras.layers.Dense(64, activation=sin))
+# model_sin.add(tf.keras.layers.Dense(64, activation=sin))
+# model_sin.add(tf.keras.layers.Dense(64, activation=sin))
+model_sin.add(tf.keras.layers.Dense(512, activation=sin))
 model_sin.add(tf.keras.layers.Dense(1))
 
 opti=tf.keras.optimizers.Adam()

code/fonctions_activations_classiques/snake.py

 import numpy as np
 
 w=10
-n=20000
+n=2000
 #création de la base de donnéé
 X,Y=creation_sin(-2.5,-1,n,w)
-X2,Y2=creation_sin(1,1.5,n,w)
+X2,Y2=creation_sin(3,3.5,n,w)
 X=np.concatenate([X,X2])
 Y=np.concatenate([Y,Y2])
 
 
 
 
+
 model_sin=tf.keras.models.Sequential()
 
 model_sin.add(tf.keras.Input(shape=(1,)))

code/fonctions_activations_classiques/swish.py

 import numpy as np
 
 w=10
-n=2000
+n=20000
 #création de la base de donnéé
-X,Y=creation_sin(-1.5,-1,n,w)
-X2,Y2=creation_sin(1,1.5,n,w)
+X,Y=creation_x2(-2.5,-1,n)
+X2,Y2=creation_x2(1,1.5,n)
 X=np.concatenate([X,X2])
 Y=np.concatenate([Y,Y2])
 
 n=10000
-Xv,Yv=creation_sin(-3,3,n,w)
-
+Xv,Yv=creation_x2(-3,3,n)
 
 
 

code/fonctions_activations_classiques/tanh_vs_ReLU.py

 from Creation_donnee import *
 import numpy as np
 w=10
-n=2000
+n=20000
 #création de la base de donnéé
-X,Y=creation_sin(-1.5,-1,n,w)
-X2,Y2=creation_sin(1,1.5,n,w)
+X,Y=creation_x2(-2.5,-1,n)
+X2,Y2=creation_x2(1,1.5,n)
 X=np.concatenate([X,X2])
 Y=np.concatenate([Y,Y2])
 
 n=10000
-Xv,Yv=creation_sin(-3,3,n,w)
+Xv,Yv=creation_x2(-3,3,n)
 
 
 
 
 
 
-model_ReLU.fit(X, Y, batch_size=16, epochs=50, shuffle='True',validation_data=(Xv, Yv))
+model_ReLU.fit(X, Y, batch_size=16, epochs=5, shuffle='True',validation_data=(Xv, Yv))
 
 
 

code/fonctions_activations_classiques/x_sin.py

 import numpy as np
 
 w=10
-n=20
+n=20000
 #création de la base de donnéé
-X,Y=creation_sin(-1.5,-1,n,w)
-X2,Y2=creation_sin(1,1.5,n,w)
+X,Y=creation_x2(-2.5,-1,n)
+X2,Y2=creation_x2(1,1.5,n)
 X=np.concatenate([X,X2])
 Y=np.concatenate([Y,Y2])
 
 n=10000
-Xv,Yv=creation_sin(-3,3,n,w)
+Xv,Yv=creation_x2(-3,3,n)
 
 
 
 
 model_xsin.summary()
 
-model_xsin.fit(X, Y, batch_size=1, epochs=10, shuffle='True',validation_data=(Xv, Yv))
+model_xsin.fit(X, Y, batch_size=16, epochs=100, shuffle='True',validation_data=(Xv, Yv))