mlp_tf_v3.py [draft+but not learning+unnec]

import glob
import numpy as np
from PIL import Image
import tensorflow as tf
import numpy as np
import matplotlib.image as mpimg
import matplotlib.pyplot as plt
from sklearn.model_selection import train_test_split
from sklearn.model_selection import KFold
from sklearn.model_selection import LeaveOneOut


img_dir_path = 'train_test_data/*.jpg'


train_test_filelist = glob.glob(img_dir_path)

#y_train_test_list = [0 if 'Y' in files else 1 for files in train_test_filelist]
X_train_test = np.array([np.array(Image.open(fname)) for fname in train_test_filelist])


learning_rate = 0.01
num_steps = 500
batch_size = 128

n_hidden_1 = 6000 # 1st layer number of neurons
n_hidden_2 = 600 # 2nd layer number of neurons
num_input = 34560 # MNIST data input (img shape: 28*28)
num_classes = 2 # MNIST total classes (0-9 digits)

def weight_variable(shape):
    initial = tf.truncated_normal(shape, stddev=0.1)
    return tf.Variable(initial)

def bias_variable(shape):
    initial = tf.constant(0.1, shape=shape)
    return tf.Variable(initial)

def init_weights(shape):
    return tf.Variable(tf.random_normal(shape, stddev=0.01))

def model(X, w_h, w_o):
    h = tf.nn.sigmoid(tf.matmul(X, w_h)) # this is a basic mlp, think 2 stacked logistic regressions
    return tf.matmul(h, w_o)

def neural_net(x):
    # Hidden fully connected layer with 256 neurons
    layer_1 = tf.add(tf.matmul(x, weights['h1']), biases['b1'])
    # Hidden fully connected layer with 256 neurons
    layer_2 = tf.add(tf.matmul(layer_1, weights['h2']), biases['b2'])
    # Output fully connected layer with a neuron for each class
    out_layer = tf.matmul(layer_2, weights['out']) + biases['out']
    return out_layer

weights = {
    'h1': tf.Variable(tf.random_normal([num_input, n_hidden_1])),
    'h2': tf.Variable(tf.random_normal([n_hidden_1, n_hidden_2])),
    'out': tf.Variable(tf.random_normal([n_hidden_2, num_classes]))
}
biases = {
    'b1': tf.Variable(tf.random_normal([n_hidden_1])),
    'b2': tf.Variable(tf.random_normal([n_hidden_2])),
    'out': tf.Variable(tf.random_normal([num_classes]))
}

array = []
with open('train_test_data_labels.txt') as fileobj:
    for line in fileobj:
        for ch in line:
            array.append(ch)

y_train_test_list = [a for a in array if a != '\n']
y_train_test = np.asarray(y_train_test_list)


X = tf.placeholder("float", [None, num_input])
#Y = tf.placeholder("float", [None, 2])
Y = tf.placeholder("float", [None, num_classes])


w_h = init_weights([34560, 10000]) # create symbolic variables
w_o = init_weights([10000, 2])

py_x = model(X, w_h, w_o)


logits = neural_net(X)
prediction = tf.nn.softmax(logits)

X_train, X_test = X_train_test[:7950], X_train_test[7950:9940]
y_train, y_test = y_train_test[:7950], y_train_test[7950:9940]

#X_train, y_train = unison_shuffled_copies(X_train, y_train)
#X_test, y_test = unison_shuffled_copies(X_test, y_test)

X_train_reshaped = X_train.reshape(7950,34560)
X_test_reshaped = X_test.reshape(1990,34560)

X_train = np.expand_dims(X_train, axis=4)###############################################################3

keep_prob = tf.placeholder(tf.float32)

train_len = len(y_train)
test_len = len(y_test)


cross_entropy = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(labels=Y, logits=logits))
#optimizer = tf.train.AdamOptimizer(learning_rate=learning_rate)
#train_op = optimizer.minimize(loss_op)

train_step = tf.train.AdamOptimizer(learning_rate=learning_rate).minimize(cross_entropy) # construct an optimizer

correct_prediction = tf.equal(tf.argmax(prediction, 1), tf.argmax(Y, 1))
accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))

config = tf.ConfigProto()
config.gpu_options.allow_growth=True

onehot_label_data_train = tf.one_hot(y_train, 2)
onehot_label_data_test = tf.one_hot(y_test, 2)

batch_size = 128

with tf.Session(config=config) as sess:
    onehot_label_data_train = sess.run(onehot_label_data_train)
    onehot_label_data_test = sess.run(onehot_label_data_test)
    sess.run(tf.global_variables_initializer())
    train_batch_count = 0
    test_batch_count = 0
    tot_train_acc = 0
    tot_test_acc = 0

    for i in range(50):
        for j in range(0,train_len,batch_size):
            train_batch_count += 1
            train_step.run(feed_dict={X:X_train_reshaped[j:j+batch_size],Y:onehot_label_data_train[j:j+batch_size],keep_prob: 0.5})
            train_accuracy = sess.run(accuracy, feed_dict={X:X_train_reshaped[j:j+batch_size,:],Y:onehot_label_data_train[j:j+batch_size,:],keep_prob: 1.0})
            tot_train_acc = tot_train_acc + train_accuracy
        avg_train_acc = tot_train_acc / (train_batch_count+1)
        test_accuracy = sess.run(accuracy, feed_dict={X:X_test_reshaped,Y:onehot_label_data_test, keep_prob: 1.0})

        print('Epoch:%g Test accuracy %g' % (i, test_accuracy))