notebook.community

Edit and run

Import





In [1]:



    


import tensorflow as tf
# from PIL import Image
import numpy as np
from scipy.misc import imread, imresize
from imagenet_classes import class_names
import os

file_path_info





In [2]:



    


#File Path
filename_queue_description = tf.train.string_input_producer(['./data/description/v2/coded_data.csv'])
filepath_ckpt  = "./ckpt/model_weight_v2_2.ckpt" #weight saver check point file path
filepath_pred = "./output/predicted_v2.csv" #predicted value file path
num_record = 50

Hyper Params - IMG





In [3]:



    


bilinear_size = 28
resized_size = bilinear_size*bilinear_size*3
img_label_size = 5

Hyper Params - LSTM





In [4]:



    


label_vec_size = 32
input_vec_size = 32
batch_size = num_record
state_size_1 = 20
state_size_2 = 100
hidden = 18
learning_rate = 0.001

vgg16_class





In [5]:



    


class vgg16:
    def __init__(self, imgs, weights=None, sess=None):
        self.imgs = imgs
        self.convlayers()
        self.fc_layers()
        self.probs = tf.nn.softmax(self.fc3l)
        if weights is not None and sess is not None:
            self.load_weights(weights, sess)


    def convlayers(self):
        self.parameters = []

        # zero-mean input
        with tf.name_scope('preprocess') as scope:
            mean = tf.constant([123.68, 116.779, 103.939], dtype=tf.float32, shape=[1, 1, 1, 3], name='img_mean')
            images = self.imgs-mean

        # conv1_1
        with tf.name_scope('conv1_1') as scope:
            kernel = tf.Variable(tf.truncated_normal([3, 3, 3, 64], dtype=tf.float32,
                                                     stddev=1e-1), name='weights')
            conv = tf.nn.conv2d(images, kernel, [1, 1, 1, 1], padding='SAME')
            biases = tf.Variable(tf.constant(0.0, shape=[64], dtype=tf.float32),
                                 trainable=True, name='biases')
            out = tf.nn.bias_add(conv, biases)
            self.conv1_1 = tf.nn.relu(out, name=scope)
            self.parameters += [kernel, biases]

        # conv1_2
        with tf.name_scope('conv1_2') as scope:
            kernel = tf.Variable(tf.truncated_normal([3, 3, 64, 64], dtype=tf.float32,
                                                     stddev=1e-1), name='weights')
            conv = tf.nn.conv2d(self.conv1_1, kernel, [1, 1, 1, 1], padding='SAME')
            biases = tf.Variable(tf.constant(0.0, shape=[64], dtype=tf.float32),
                                 trainable=True, name='biases')
            out = tf.nn.bias_add(conv, biases)
            self.conv1_2 = tf.nn.relu(out, name=scope)
            self.parameters += [kernel, biases]

        # pool1
        self.pool1 = tf.nn.max_pool(self.conv1_2,
                               ksize=[1, 2, 2, 1],
                               strides=[1, 2, 2, 1],
                               padding='SAME',
                               name='pool1')

        # conv2_1
        with tf.name_scope('conv2_1') as scope:
            kernel = tf.Variable(tf.truncated_normal([3, 3, 64, 128], dtype=tf.float32,
                                                     stddev=1e-1), name='weights')
            conv = tf.nn.conv2d(self.pool1, kernel, [1, 1, 1, 1], padding='SAME')
            biases = tf.Variable(tf.constant(0.0, shape=[128], dtype=tf.float32),
                                 trainable=True, name='biases')
            out = tf.nn.bias_add(conv, biases)
            self.conv2_1 = tf.nn.relu(out, name=scope)
            self.parameters += [kernel, biases]

        # conv2_2
        with tf.name_scope('conv2_2') as scope:
            kernel = tf.Variable(tf.truncated_normal([3, 3, 128, 128], dtype=tf.float32,
                                                     stddev=1e-1), name='weights')
            conv = tf.nn.conv2d(self.conv2_1, kernel, [1, 1, 1, 1], padding='SAME')
            biases = tf.Variable(tf.constant(0.0, shape=[128], dtype=tf.float32),
                                 trainable=True, name='biases')
            out = tf.nn.bias_add(conv, biases)
            self.conv2_2 = tf.nn.relu(out, name=scope)
            self.parameters += [kernel, biases]

        # pool2
        self.pool2 = tf.nn.max_pool(self.conv2_2,
                               ksize=[1, 2, 2, 1],
                               strides=[1, 2, 2, 1],
                               padding='SAME',
                               name='pool2')

        # conv3_1
        with tf.name_scope('conv3_1') as scope:
            kernel = tf.Variable(tf.truncated_normal([3, 3, 128, 256], dtype=tf.float32,
                                                     stddev=1e-1), name='weights')
            conv = tf.nn.conv2d(self.pool2, kernel, [1, 1, 1, 1], padding='SAME')
            biases = tf.Variable(tf.constant(0.0, shape=[256], dtype=tf.float32),
                                 trainable=True, name='biases')
            out = tf.nn.bias_add(conv, biases)
            self.conv3_1 = tf.nn.relu(out, name=scope)
            self.parameters += [kernel, biases]

        # conv3_2
        with tf.name_scope('conv3_2') as scope:
            kernel = tf.Variable(tf.truncated_normal([3, 3, 256, 256], dtype=tf.float32,
                                                     stddev=1e-1), name='weights')
            conv = tf.nn.conv2d(self.conv3_1, kernel, [1, 1, 1, 1], padding='SAME')
            biases = tf.Variable(tf.constant(0.0, shape=[256], dtype=tf.float32),
                                 trainable=True, name='biases')
            out = tf.nn.bias_add(conv, biases)
            self.conv3_2 = tf.nn.relu(out, name=scope)
            self.parameters += [kernel, biases]

        # conv3_3
        with tf.name_scope('conv3_3') as scope:
            kernel = tf.Variable(tf.truncated_normal([3, 3, 256, 256], dtype=tf.float32,
                                                     stddev=1e-1), name='weights')
            conv = tf.nn.conv2d(self.conv3_2, kernel, [1, 1, 1, 1], padding='SAME')
            biases = tf.Variable(tf.constant(0.0, shape=[256], dtype=tf.float32),
                                 trainable=True, name='biases')
            out = tf.nn.bias_add(conv, biases)
            self.conv3_3 = tf.nn.relu(out, name=scope)
            self.parameters += [kernel, biases]

        # pool3
        self.pool3 = tf.nn.max_pool(self.conv3_3,
                               ksize=[1, 2, 2, 1],
                               strides=[1, 2, 2, 1],
                               padding='SAME',
                               name='pool3')

        # conv4_1
        with tf.name_scope('conv4_1') as scope:
            kernel = tf.Variable(tf.truncated_normal([3, 3, 256, 512], dtype=tf.float32,
                                                     stddev=1e-1), name='weights')
            conv = tf.nn.conv2d(self.pool3, kernel, [1, 1, 1, 1], padding='SAME')
            biases = tf.Variable(tf.constant(0.0, shape=[512], dtype=tf.float32),
                                 trainable=True, name='biases')
            out = tf.nn.bias_add(conv, biases)
            self.conv4_1 = tf.nn.relu(out, name=scope)
            self.parameters += [kernel, biases]

        # conv4_2
        with tf.name_scope('conv4_2') as scope:
            kernel = tf.Variable(tf.truncated_normal([3, 3, 512, 512], dtype=tf.float32,
                                                     stddev=1e-1), name='weights')
            conv = tf.nn.conv2d(self.conv4_1, kernel, [1, 1, 1, 1], padding='SAME')
            biases = tf.Variable(tf.constant(0.0, shape=[512], dtype=tf.float32),
                                 trainable=True, name='biases')
            out = tf.nn.bias_add(conv, biases)
            self.conv4_2 = tf.nn.relu(out, name=scope)
            self.parameters += [kernel, biases]

        # conv4_3
        with tf.name_scope('conv4_3') as scope:
            kernel = tf.Variable(tf.truncated_normal([3, 3, 512, 512], dtype=tf.float32,
                                                     stddev=1e-1), name='weights')
            conv = tf.nn.conv2d(self.conv4_2, kernel, [1, 1, 1, 1], padding='SAME')
            biases = tf.Variable(tf.constant(0.0, shape=[512], dtype=tf.float32),
                                 trainable=True, name='biases')
            out = tf.nn.bias_add(conv, biases)
            self.conv4_3 = tf.nn.relu(out, name=scope)
            self.parameters += [kernel, biases]

        # pool4
        self.pool4 = tf.nn.max_pool(self.conv4_3,
                               ksize=[1, 2, 2, 1],
                               strides=[1, 2, 2, 1],
                               padding='SAME',
                               name='pool4')

        # conv5_1
        with tf.name_scope('conv5_1') as scope:
            kernel = tf.Variable(tf.truncated_normal([3, 3, 512, 512], dtype=tf.float32,
                                                     stddev=1e-1), name='weights')
            conv = tf.nn.conv2d(self.pool4, kernel, [1, 1, 1, 1], padding='SAME')
            biases = tf.Variable(tf.constant(0.0, shape=[512], dtype=tf.float32),
                                 trainable=True, name='biases')
            out = tf.nn.bias_add(conv, biases)
            self.conv5_1 = tf.nn.relu(out, name=scope)
            self.parameters += [kernel, biases]

        # conv5_2
        with tf.name_scope('conv5_2') as scope:
            kernel = tf.Variable(tf.truncated_normal([3, 3, 512, 512], dtype=tf.float32,
                                                     stddev=1e-1), name='weights')
            conv = tf.nn.conv2d(self.conv5_1, kernel, [1, 1, 1, 1], padding='SAME')
            biases = tf.Variable(tf.constant(0.0, shape=[512], dtype=tf.float32),
                                 trainable=True, name='biases')
            out = tf.nn.bias_add(conv, biases)
            self.conv5_2 = tf.nn.relu(out, name=scope)
            self.parameters += [kernel, biases]

        # conv5_3
        with tf.name_scope('conv5_3') as scope:
            kernel = tf.Variable(tf.truncated_normal([3, 3, 512, 512], dtype=tf.float32,
                                                     stddev=1e-1), name='weights')
            conv = tf.nn.conv2d(self.conv5_2, kernel, [1, 1, 1, 1], padding='SAME')
            biases = tf.Variable(tf.constant(0.0, shape=[512], dtype=tf.float32),
                                 trainable=True, name='biases')
            out = tf.nn.bias_add(conv, biases)
            self.conv5_3 = tf.nn.relu(out, name=scope)
            self.parameters += [kernel, biases]

        # pool5
        self.pool5 = tf.nn.max_pool(self.conv5_3,
                               ksize=[1, 2, 2, 1],
                               strides=[1, 2, 2, 1],
                               padding='SAME',
                               name='pool4')

    def fc_layers(self):
        # fc1
        with tf.name_scope('fc1') as scope:
            shape = int(np.prod(self.pool5.get_shape()[1:]))
            fc1w = tf.Variable(tf.truncated_normal([shape, 4096],
                                                         dtype=tf.float32,
                                                         stddev=1e-1), name='weights')
            fc1b = tf.Variable(tf.constant(1.0, shape=[4096], dtype=tf.float32),
                                 trainable=True, name='biases')
            pool5_flat = tf.reshape(self.pool5, [-1, shape])
            fc1l = tf.nn.bias_add(tf.matmul(pool5_flat, fc1w), fc1b)
            self.fc1 = tf.nn.relu(fc1l)
            self.parameters += [fc1w, fc1b]

        # fc2
        with tf.name_scope('fc2') as scope:
            fc2w = tf.Variable(tf.truncated_normal([4096, 4096],
                                                         dtype=tf.float32,
                                                         stddev=1e-1), name='weights')
            fc2b = tf.Variable(tf.constant(1.0, shape=[4096], dtype=tf.float32),
                                 trainable=True, name='biases')
            fc2l = tf.nn.bias_add(tf.matmul(self.fc1, fc2w), fc2b)
            self.fc2 = tf.nn.relu(fc2l)
            self.parameters += [fc2w, fc2b]

        # fc3
        with tf.name_scope('fc3') as scope:
            fc3w = tf.Variable(tf.truncated_normal([4096, 1000],
                                                         dtype=tf.float32,
                                                         stddev=1e-1), name='weights')
            fc3b = tf.Variable(tf.constant(1.0, shape=[1000], dtype=tf.float32),
                                 trainable=True, name='biases')
            self.fc3l = tf.nn.bias_add(tf.matmul(self.fc2, fc3w), fc3b)
            self.parameters += [fc3w, fc3b]

    def load_weights(self, weight_file, sess):
        weights = np.load(weight_file)
        keys = sorted(weights.keys())
        for i, k in enumerate(keys):
            print(i, k, np.shape(weights[k]))
            sess.run(self.parameters[i].assign(weights[k]))

Train

load_img_vgg16





In [ ]:



    


with tf.Session() as sess_vgg:
    imgs = tf.placeholder(tf.float32, [None, 200, 200, 3])
    vgg = vgg16(imgs, 'vgg16_weights.npz', sess_vgg)
    img_files = ['./data/img/cropped/' + i for i in os.listdir('./data/img/cropped')]
    imgs = [imread(file, mode='RGB') for file in img_files]
    #bilinear feature
    imgs_bi = [sess_vgg.run(vgg.fc1, feed_dict={vgg.imgs: [img]})[0] for img in imgs]
#     imgs_bi = [imresize(arr=img, interp='bilinear', size=bilinear_size) for img in imgs]
    imgs_bi = np.reshape(a=imgs_bi, newshape=[50,-1])
    
    #label
    prob = [sess_vgg.run(vgg.probs, feed_dict={vgg.imgs: [img]})[0] for img in imgs]
    preds = [(np.argsort(p)[::-1])[0:1] for p in prob]
    preds = [p[0] for p in preds]
    for i in range(len(preds)):
        if(preds[i]==430):
            preds[i]=0
        elif(preds[i]==429):
            preds[i]=1
        elif(preds[i]==805):
            preds[i]=2
        elif(preds[i]==768):
            preds[i]=3
        elif(preds[i]==574):
            preds[i]=4
    
    img_label_onehot = tf.one_hot(indices=preds, depth = 5)
    print(preds)
#     print(sess_vgg.run(img_label_onehot))
    #clear
    imgs = None
    vgg = None
    img_files = None
    prob = None
    sess_vgg.close()



    











In [ ]:



    


print(np.shape(imgs_bi))

Text_Reader_Setting





In [ ]:



    


reader = tf.TextLineReader()
key,value = reader.read(filename_queue_description)
record_defaults =[[-2], [-2], [-2], [-2], [-2], [-2], [-2], [-2], [-2], [-2], [-2], [-2], [-2], [-2], [-2], [-2], [-2], [-2], [-2]]
w1, w2, w3, w4, w5, w6, w7, w8, w9, w10, w11, w12, w13, w14, w15, w16, w17, w18, w19 = tf.decode_csv(value, record_defaults)  

feature_label = tf.stack([w2, w3, w4, w5, w6, w7, w8, w9, w10, w11, w12, w13, w14, w15, w16, w17, w18, w19])
feature_word = tf.stack([w1, w2, w3, w4, w5, w6, w7, w8, w9, w10, w11, w12, w13, w14, w15, w16, w17, w18])



    











In [ ]:



    


with tf.Session() as sess_data:
    coord = tf.train.Coordinator()
    threads = tf.train.start_queue_runners(coord=coord)
#     img_queue = []
    for i in range(num_record):
#         image = sess.run(images)
        
        raw_label, raw_input = sess_data.run([feature_label, feature_word])
        onehot_input = tf.one_hot(indices=raw_input, depth=32)
        onehot_label = tf.one_hot(indices=raw_label, depth=32)
        if i == 0:
            full_input = onehot_input
            full_label = onehot_label
        else:
            full_input = tf.concat([full_input, onehot_input], 0)
            full_label = tf.concat([full_label, onehot_label], 0)
        
    raw_label = None
    raw_input = None
    onehot_input = None
    onehot_label = None
    coord.request_stop()
    coord.join(threads)
    sess_data.close()



    











In [ ]:



    


with tf.name_scope('batch') as scope:
    # full_label = tf.reshape(full_label, [batch_size, hidden, label_vec_size])
    full_input = tf.reshape(full_input, [batch_size, hidden, input_vec_size])
    full_label = tf.reshape(full_label, [batch_size, hidden, input_vec_size])
    
#     input_batch, label_batch = tf.train.batch([full_input, full_label], batch_size=50)

Model

LSTM First Layer





In [ ]:



    


with tf.name_scope('lstm_layer_1') as scope:
    with tf.variable_scope('lstm_layer_1'):
        rnn_cell_1 = tf.contrib.rnn.BasicLSTMCell(state_size_1, reuse=None)
        output_1, _ = tf.contrib.rnn.static_rnn(rnn_cell_1, tf.unstack(full_input, axis=1), dtype=tf.float32)
        input_2 = [tf.concat([out, imgs_bi, img_label_onehot], axis=1) for out in output_1]
#         output_w_1 = tf.Variable(tf.truncated_normal([hidden, state_size_1, input_vec_size]))
#         output_b_1 = tf.Variable(tf.zeros([input_vec_size]))
#         pred_temp = tf.matmul(output_1, output_w_1) + output_b_1

LSTM Second Layer





In [ ]:



    


with tf.name_scope('lstm_layer_2') as scope:
    with tf.variable_scope('lstm_layer_2'):
        rnn_cell_2 = tf.contrib.rnn.BasicLSTMCell(state_size_2, reuse=None)
        output_2, _ = tf.contrib.rnn.static_rnn(rnn_cell_2, tf.unstack(input_2, axis=0), dtype=tf.float32)
        output_w_2 = tf.Variable(tf.truncated_normal([hidden, state_size_2, input_vec_size]))
        output_b_2 = tf.Variable(tf.zeros([input_vec_size]))
        pred = tf.nn.softmax(tf.matmul(output_2, output_w_2) + output_b_2)

Loss





In [ ]:



    


with tf.name_scope('loss') as scope:
    loss = tf.constant(0, tf.float32)
    for i in range(hidden):
        loss += tf.losses.softmax_cross_entropy(tf.unstack(full_label, axis=1)[i], tf.unstack(pred, axis=0)[i])
    train = tf.train.AdamOptimizer(learning_rate).minimize(loss)

Run Train





In [ ]:



    


sess_train = tf.Session()
sess_train.run(tf.global_variables_initializer())
saver = tf.train.Saver()
save_path = saver.save(sess_train, filepath_ckpt)



    











In [ ]:



    


for i in range(10001):
    sess_train.run(train)
    if i % 1000 == 0:
        print("loss : ", sess_train.run(loss))
#             print("pred : ", sess.run(pred))
save_path = saver.save(sess_train, filepath_ckpt)
print("= Weigths are saved in " + filepath_ckpt)



    











In [ ]:



    


sess_train.close()

Test





In [6]:



    


with tf.Session() as sess_vgg_test:
    imgs = tf.placeholder(tf.float32, [None, 200, 200, 3])
    vgg = vgg16(imgs, 'vgg16_weights.npz', sess_vgg_test)
    test_img_files = ['./data/img/test/cropped/test001.png', './data/img/cropped/005.png', './data/img/cropped/014.png', './data/img/cropped/021.png', './data/img/cropped/036.png', './data/img/cropped/050.png']
    num_imgs = len(test_img_files)
    test_imgs = [imread(file, mode='RGB') for file in test_img_files]
    
    #bilinear feature
    test_imgs_bi = [sess_vgg_test.run(vgg.fc1, feed_dict={vgg.imgs: [img]})[0] for img in test_imgs]
#     test_imgs_bi = [imresize(arr=img, interp='bilinear', size=bilinear_size) for img in test_imgs]
    test_imgs_bi = np.reshape(a=test_imgs_bi, newshape=[num_imgs,-1])
 
    #label
    prob = [sess_vgg_test.run(vgg.probs, feed_dict={vgg.imgs: [img]})[0] for img in test_imgs]
    test_preds = [(np.argsort(p)[::-1])[0:1] for p in prob]
    test_preds = [p[0] for p in test_preds]
    for i in range(len(test_preds)):
        if(test_preds[i]==430):
            test_preds[i]=0
        elif(test_preds[i]==429):
            test_preds[i]=1
        elif(test_preds[i]==805):
            test_preds[i]=2
        elif(test_preds[i]==768):
            test_preds[i]=3
        elif(test_preds[i]==574):
            test_preds[i]=4
    test_img_label_onehot = tf.one_hot(indices=test_preds, depth = 5)
    print(sess_vgg_test.run(test_img_label_onehot))

    #clear
    test_imgs = None
    vgg = None
    test_img_files = None
    prob = None
    
    sess_vgg_test.close()



    


















    






0 conv1_1_W (3, 3, 3, 64)
1 conv1_1_b (64,)
2 conv1_2_W (3, 3, 64, 64)
3 conv1_2_b (64,)
4 conv2_1_W (3, 3, 64, 128)
5 conv2_1_b (128,)
6 conv2_2_W (3, 3, 128, 128)
7 conv2_2_b (128,)
8 conv3_1_W (3, 3, 128, 256)
9 conv3_1_b (256,)
10 conv3_2_W (3, 3, 256, 256)
11 conv3_2_b (256,)
12 conv3_3_W (3, 3, 256, 256)
13 conv3_3_b (256,)
14 conv4_1_W (3, 3, 256, 512)
15 conv4_1_b (512,)
16 conv4_2_W (3, 3, 512, 512)
17 conv4_2_b (512,)
18 conv4_3_W (3, 3, 512, 512)
19 conv4_3_b (512,)
20 conv5_1_W (3, 3, 512, 512)
21 conv5_1_b (512,)
22 conv5_2_W (3, 3, 512, 512)
23 conv5_2_b (512,)
24 conv5_3_W (3, 3, 512, 512)
25 conv5_3_b (512,)
26 fc6_W (25088, 4096)
27 fc6_b (4096,)
28 fc7_W (4096, 4096)
29 fc7_b (4096,)
30 fc8_W (4096, 1000)
31 fc8_b (1000,)
[[ 1.  0.  0.  0.  0.]
 [ 1.  0.  0.  0.  0.]
 [ 0.  1.  0.  0.  0.]
 [ 0.  0.  1.  0.  0.]
 [ 0.  0.  0.  1.  0.]
 [ 0.  0.  0.  0.  1.]]

















In [7]:



    


start_input = tf.zeros([num_imgs,hidden,input_vec_size])
with tf.Session() as sess_init_generator:
    input_init = sess_init_generator.run(start_input)
for i in range(num_imgs):
    input_init[i][0][0] = 1

Test-First_Layer-LSTM





In [8]:



    


with tf.name_scope('lstm_layer_1') as scope:
    with tf.variable_scope('lstm_layer_1'):
        rnn_cell_1 = tf.contrib.rnn.BasicLSTMCell(state_size_1, reuse=None)
        output_test_1, _ = tf.contrib.rnn.static_rnn(rnn_cell_1, tf.unstack(input_init, axis=1), dtype=tf.float32)
        input_2 = [tf.concat([out, test_imgs_bi, test_img_label_onehot], axis=1) for out in output_test_1]
# output_t_1 = tf.contrib.rnn.static_rnn(rnn_cell, tf.unstack(full_input, axis=1), dtype=tf.float32)
# pred = tf.nn.softmax(tf.matmul(output1, output_w[0]) + output_b[0])



    











In [9]:



    


with tf.name_scope('lstm_layer_2') as scope:
    with tf.variable_scope('lstm_layer_2'):
        rnn_cell_2 = tf.contrib.rnn.BasicLSTMCell(state_size_2, reuse=None)
        output_2, _ = tf.contrib.rnn.static_rnn(rnn_cell_2, tf.unstack(input_2, axis=0), dtype=tf.float32)
        output_w_2 = tf.Variable(tf.truncated_normal([hidden, state_size_2, input_vec_size]))
        output_b_2 = tf.Variable(tf.zeros([input_vec_size]))
        pred = tf.nn.softmax(tf.matmul(output_2, output_w_2) + output_b_2)



    











In [10]:



    


sess_model = tf.Session()
saver = tf.train.Saver(allow_empty=True)
saver.restore(sess_model, filepath_ckpt)



    


















    






INFO:tensorflow:Restoring parameters from ./ckpt/model_weight_v2_2.ckpt
first_test = sess_model.run(pred)[0] input_init[0][1] = first_test
second_test = sess_model.run(pred)[1] input_init[0][2] = second_test
third_test = sess_model.run(pred)[2] input_init[0][3] = third_test 




In [11]:



    


for i in range(hidden):
    result = sess_model.run(pred)
    result_temp = result[i]
    if i == hidden -1:
        pass
    else:
        input_init[:,i+1] = result_temp



    











In [12]:



    


print(np.shape(result))



    


















    






(18, 6, 32)

Result Check





In [13]:



    


import csv
classes = []
f = open('./data/description/v2/class.csv', 'r')
csvReader = csv.reader(f)
for row in csvReader:
    classes.append(row)
#     print(row)
f.close()



    











In [14]:



    


decoded_result = np.argmax(a=result, axis=2)
print(np.shape(decoded_result))



    


















    






(18, 6)

















In [18]:



    


for i in range(num_imgs):
    str = " "
    for r in decoded_result:
        str += classes[r[i]][0] + " "
    print(str)



    


















    






 this is basketball because this is round and brown color with black pattern lines <EOS> <EOS> <EOS> <EOS> 
 this is basketball because this is round and brown color with black pattern lines <EOS> <EOS> <EOS> <EOS> 
 this is baseball because this is round and white color with red pattern lines <EOS> <EOS> <EOS> <EOS> 
 this is soccerball  because this is round and white color with black pattern lines <EOS> <EOS> <EOS> <EOS> 
 this is lugbyball  because this is on and brown color with white pattern lines <EOS> <EOS> <EOS> <EOS> 
 this is golfball because this is round and white color with dimply pattern <EOS> <EOS> <EOS> <EOS> <EOS>

Code Storage

def lstm_cell(): tf.contrib.rnn.BasicLSTMCell(size, forget_bias=0.0, state_is_tuple=True) inputs = tf.nn.embedding_lookup(embedding, input_.input_data) if is_training and config.keep_prob < 1: output = tf.reshape(tf.stack(axis=1, values=outputs), [-1, size]) softmax_w = tf.get_variable("softmax_w", [size, vocab_size], dtype=data_type()) softmax_b = tf.get_variable("softmax_b", [vocab_size], dtype=data_type()) logits = tf.matmul(output, softmax_w) + softmax_b loss = tf.contrib.legacy_seq2seq.sequence_loss_by_example( [logits], [tf.reshape(input_.targets, [-1])], [tf.ones([batch_size * num_steps], dtype=data_type())]) self._cost = cost = tf.reduce_sum(loss) / batch_size self._final_state = state
sess = tf.Session() imgs = tf.placeholder(tf.float32, [None, 200, 200, 3]) vgg = vgg16(imgs, 'vgg16_weights.npz', sess) img1 = imread('data/img/cropped/002.png', mode='RGB') prob = sess.run(vgg.probs, feed_dict={vgg.imgs: [img1]})[0] preds = (np.argsort(prob)[::-1])[0:5] for p in preds: print(class_names[p], prob[p])

Content source: Timmy-Oh/Generating-Visual-Explanation

Similar notebooks:

notebook.community | gallery | about