Generating fractals

import tensorflow as tf import matplotlib.pyplot as plt import numpy as np import pandas as pd import datetime from google.colab import drive print(tf.__version__) # Generate a learning set class MandelbrotDataSet: def __init__(self, size=256, max_depth=100, xmin=-2.0, xmax=0.7, ymin=-1.3, ymax=1.3): self.data = {} self.x = tf.random.uniform((size,),xmin,xmax,tf.float32) self.y = tf.random.uniform((size,),ymin,ymax,tf.float32) self.outputs = self.mandel(x=self.x, y=self.y,max_depth=max_depth) self.data = tf.stack([self.x, self.y], axis=1) @staticmethod def mandel(x, y, max_depth): zx, zy = x,y for n in range(1, max_depth): zx, zy = zx*zx - zy*zy + x, 2*zx*zy + y return tf.less(zx*zx+zy*zy, 4.0) class MandelbrotPointGenerator(tf.keras.utils.Sequence): def __init__(self): self.x = {} self.y = {} def __len__(self): return 128 def __getitem__(self, epoch): self.x = tf.random.uniform((4096,),-2.0,0.7,tf.float32) self.y = tf.random.uniform((4096,),-1.3,1.3,tf.float32) return tf.stack([self.x, self.y], axis=1), self.mandel(x=self.x, y=self.y,max_depth=100) @staticmethod def mandel(x, y, max_depth): zx, zy = x,y for n in range(1, max_depth): zx, zy = zx*zx - zy*zy + x, 2*zx*zy + y return tf.less(zx*zx+zy*zy, 4.0) #trainingData = MandelbrotDataSet(10_000) model = tf.keras.Sequential([ tf.keras.Input(shape=(2,)), tf.keras.layers.Dense(1024,activation="LeakyReLU"), tf.keras.layers.Dense(1024,activation="LeakyReLU"), tf.keras.layers.Dense(512,activation="LeakyReLU"), tf.keras.layers.Dense(256,activation="LeakyReLU"), tf.keras.layers.Dense(128,activation="LeakyReLU"), tf.keras.layers.Dense(64,activation="LeakyReLU"), tf.keras.layers.Dense(32,activation="LeakyReLU"), tf.keras.layers.Dense(8,activation="LeakyReLU"), tf.keras.layers.Dense(8,activation="LeakyReLU"), tf.keras.layers.Dense(2,activation="LeakyReLU"), tf.keras.layers.Dense(1,activation="sigmoid") ]) #model.compile(loss=tf.keras.losses.MeanSquaredError(), optimizer=tf.keras.optimizers.Adam(learning_rate=0.001),metrics=["mae"]) model.compile(optimizer="adam", loss="binary_crossentropy", metrics=["accuracy"]) log_dir = "logs/fit/" + datetime.datetime.now().strftime("%Y%m%d-%H%M%S") tensorboard_callback = tf.keras.callbacks.TensorBoard(log_dir=log_dir, histogram_freq=1) sequence = MandelbrotPointGenerator() history = model.fit(sequence,epochs=20) np.set_printoptions(precision=3, suppress=True) hist = pd.DataFrame(history.history) hist['epoch'] = history.epoch hist.tail() def plot_loss(history): plt.plot(history.history['accuracy'], label='accuracy') plt.xlabel('Epoch') plt.ylabel('accuracy') plt.legend() plt.grid(True) plot_loss(history) trainingData = MandelbrotDataSet(100_000) predictions = model.predict(trainingData.data) print("predictions shape:", predictions.shape) plt.scatter(trainingData.x, trainingData.y, s=1, c=predictions) model.save('my_model') new_model = tf.keras.models.load_model('my_model') # Check its architecture new_model.summary() trainingData = MandelbrotDataSet(size=100_000, max_depth=100, xmin=-2.0, xmax=-1.0, ymin=-0.3, ymax=0.3) new_predictions = new_model.predict(trainingData.data) print("predictions shape:", predictions.shape) plt.scatter(trainingData.x, trainingData.y, s=1, c=new_predictions) # Mount Google Drive #drive.mount('/content/drive') # Create a folder in the root directory #model.save('/content/drive/My Drive/My Model/dynamic_tensorbroat_3')

import tensorflow as tf import matplotlib.pyplot as plt import numpy as np import pandas as pd import datetime from google.colab import drive print(tf.__version__) drive.mount('/content/drive', force_remount=False) new_model = tf.keras.models.load_model('/content/drive/My Drive/My Model/dynamic_tensorbroat_3') # Check its architecture new_model.summary() class MandelbrotPointGenerator(tf.keras.utils.Sequence): def __init__(self): self.x = {} self.y = {} def __len__(self): return 512 def __getitem__(self, epoch): self.x = tf.random.uniform((512,),-2.0,0.7,tf.float32) self.y = tf.random.uniform((512,),-1.3,1.3,tf.float32) return tf.stack([self.x, self.y], axis=1), self.mandel(x=self.x, y=self.y,max_depth=100) @staticmethod def mandel(x, y, max_depth): zx, zy = x,y for n in range(1, max_depth): zx, zy = zx*zx - zy*zy + x, 2*zx*zy + y return tf.less(zx*zx+zy*zy, 4.0) sequence = MandelbrotPointGenerator() new_model.fit(sequence, epochs=200) new_model.save("/content/drive/My Drive/My Model/dynamic_tensorbroat_3") # Generate a learning set class MandelbrotDataSet: def __init__(self, size=256, max_depth=100, xmin=-2.0, xmax=0.7, ymin=-1.3, ymax=1.3): self.data = {} self.x = tf.random.uniform((size,),xmin,xmax,tf.float32) self.y = tf.random.uniform((size,),ymin,ymax,tf.float32) self.outputs = self.mandel(x=self.x, y=self.y,max_depth=max_depth) self.data = tf.stack([self.x, self.y], axis=1) @staticmethod def mandel(x, y, max_depth): zx, zy = x,y for n in range(1, max_depth): zx, zy = zx*zx - zy*zy + x, 2*zx*zy + y return tf.less(zx*zx+zy*zy, 4.0) trainingData = MandelbrotDataSet(size=200_000, max_depth=100, xmin=-2.0, xmax=0.7, ymin=-1.3, ymax=1.3) new_predictions = new_model.predict(trainingData.data) plt.scatter(trainingData.x, trainingData.y, s=1, c=new_predictions) trainingData = MandelbrotDataSet(size=200_000, max_depth=100, xmin=-2.0, xmax=-1.0, ymin=-0.3, ymax=0.3) new_predictions = new_model.predict(trainingData.data) plt.scatter(trainingData.x, trainingData.y, s=1, c=new_predictions)