08_deeplearning_python

softmax

import numpy as np
import matplotlib.pyplot as plt
plt.style.use('./deeplearning.mplstyle')
import tensorflow as tf
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense
from IPython.display import display, Markdown, Latex
from sklearn.datasets import make_blobs
%matplotlib widget
from matplotlib.widgets import Slider
from lab_utils_common import dlc
from lab_utils_softmax import plt_softmax
import logging
logging.getLogger("tensorflow").setLevel(logging.ERROR)
tf.autograph.set_verbosity(0)

def my_softmax(z):
    ez = np.exp(z)              #element-wise exponenial
    sm = ez/np.sum(ez)
    return(sm)
plt.close("all")
plt_softmax(my_softmax)
# make  dataset for example
centers = [[-5, 2], [-2, -2], [1, 2], [5, -2]]
X_train, y_train = make_blobs(n_samples=2000, centers=centers, cluster_std=1.0,random_state=30)

model = Sequential(
    [ 
        Dense(25, activation = 'relu'),
        Dense(15, activation = 'relu'),
        Dense(4, activation = 'softmax')    # < softmax activation here
    ]
)
model.compile(
    loss=tf.keras.losses.SparseCategoricalCrossentropy(),
    optimizer=tf.keras.optimizers.Adam(0.001),
)

model.fit(
    X_train,y_train,
    epochs=10
)
        

preferred_model = Sequential(
    [ 
        Dense(25, activation = 'relu'),
        Dense(15, activation = 'relu'),
        Dense(4, activation = 'linear')   #<-- Note
    ]
)
preferred_model.compile(
    loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),  #<-- Note
    optimizer=tf.keras.optimizers.Adam(0.001),
)

preferred_model.fit(
    X_train,y_train,
    epochs=10
)
        

p_nonpreferred = model.predict(X_train) #predict which is the best
print(p_nonpreferred [:2])
print("largest value", np.max(p_nonpreferred), "smallest value", np.min(p_nonpreferred))
p_preferred = preferred_model.predict(X_train)  #didn't get the probilities
print(f"two example output vectors:\n {p_preferred[:2]}")
print("largest value", np.max(p_preferred), "smallest value", np.min(p_preferred))
sm_preferred = tf.nn.softmax(p_preferred).numpy()  #to get the probilities
print(f"two example output vectors:\n {sm_preferred[:2]}")
print("largest value", np.max(sm_preferred), "smallest value", np.min(sm_preferred))
# the np.argmax():can get the largest output's index.
for i in range(5):
    print( f"{p_preferred[i]}, category: {np.argmax(p_preferred[i])}")

multi-class classification

import numpy as np
import matplotlib.pyplot as plt
%matplotlib widget
from sklearn.datasets import make_blobs
import tensorflow as tf
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense
np.set_printoptions(precision=2)
from lab_utils_multiclass_TF import *
import logging
logging.getLogger("tensorflow").setLevel(logging.ERROR)
tf.autograph.set_verbosity(0)
# make 4-class dataset for classification
classes = 4
m = 100
centers = [[-5, 2], [-2, -2], [1, 2], [5, -2]]
std = 1.0
X_train, y_train = make_blobs(n_samples=m, centers=centers, cluster_std=std,random_state=30)
plt_mc(X_train,y_train,classes, centers, std=std)

# show classes in data set
print(f"unique classes {np.unique(y_train)}")
# show how classes are represented
print(f"class representation {y_train[:10]}")
# show shapes of our dataset
print(f"shape of X_train: {X_train.shape}, shape of y_train: {y_train.shape}")
'''
unique classes [0 1 2 3]
class representation [3 3 3 0 3 3 3 3 2 0]
shape of X_train: (100, 2), shape of y_train: (100,)
'''
tf.random.set_seed(1234)  # applied to achieve consistent results
model = Sequential(
    [
        Dense(2, activation = 'relu',   name = "L1"),
        Dense(4, activation = 'linear', name = "L2")
    ]
)
model.compile(
    loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),
    optimizer=tf.keras.optimizers.Adam(0.01),
)

model.fit(
    X_train,y_train,
    epochs=200
)
plt_cat_mc(X_train, y_train, model, classes)

# gather the trained parameters from the first layer
l1 = model.get_layer("L1")
W1,b1 = l1.get_weights()
# plot the function of the first layer
plt_layer_relu(X_train, y_train.reshape(-1,), W1, b1, classes)
# gather the trained parameters from the output layer
l2 = model.get_layer("L2")
W2, b2 = l2.get_weights()
# create the 'new features', the training examples after L1 transformation
Xl2 = np.maximum(0, np.dot(X_train,W1) + b1)

plt_output_layer_linear(Xl2, y_train.reshape(-1,), W2, b2, classes,
                        x0_rng = (-0.25,np.amax(Xl2[:,0])), x1_rng = (-0.25,np.amax(Xl2[:,1])))