Related
Below is the code of what I'm trying to do, but my accuracy is always under 50% so I'm wondering how should I fix this? What I'm trying to do is use the first 1885 daily unit sale data as input and the rest of the daily unit sale data from 1885 as output. After train these data, I need to use it to predict 20 more daily unit sale in the future
The data I used here is provided in this link
https://drive.google.com/file/d/13qzIZMD6Wz7e1GpOsNw1_9Yq-4PI2HrC/view?usp=sharing
import pandas as pd
import numpy as np
import keras
import keras.backend as k
import tensorflow as tf
from keras.models import Sequential
from keras.layers import Dense
from keras.layers import Dropout
from keras.callbacks import EarlyStopping
from sklearn import preprocessing
from sklearn.preprocessing import StandardScaler
from sklearn.model_selection import train_test_split
data = pd.read_csv('sales_train.csv')
#Since there are 3 departments and 10 store from 3 different areas, thus I categorized the data into 30 groups and numerize them
Unique_dept = data["dept_id"].unique()
Unique_state = data['state_id'].unique()
Unique_store = data["store_id"].unique()
data0 = data.copy()
for i in range(3):
data0["dept_id"] = data0["dept_id"].replace(to_replace=Unique_dept[i], value = i)
data0["state_id"] = data0["state_id"].replace(to_replace=Unique_state[i], value = i)
for j in range(10):
data0["store_id"] = data0["store_id"].replace(to_replace=Unique_store[j], value = int(Unique_store[j][3]) -1)
# Select the three numerized categorical variables and daily unit sale data
pt = 6 + 1885
X = pd.concat([data0.iloc[:,2],data0.iloc[:, 4:pt]], axis = 1)
Y = data0.iloc[:, pt:]
# Remove the daily unit sale data that are highly correlated to each other (corr > 0.9)
correlation = X.corr(method = 'pearson')
corr_lst = []
for i in correlation:
for j in correlation:
if (i != j) & (correlation[i][j] >= 0.9) & (j not in corr_lst) & (i not in corr_lst):
corr_lst.append(i)
x = X.drop(corr_lst, axis = 1)
x_value = x.values
y_value = Y.values
sc = StandardScaler()
X_scale = sc.fit_transform(x_value)
X_train, X_val_and_test, Y_train, Y_val_and_test = train_test_split(x_value, y_value, test_size=0.2)
X_val, X_test, Y_val, Y_test = train_test_split(X_val_and_test, Y_val_and_test, test_size=0.5)
print(X_train.shape, X_val.shape, X_test.shape, Y_train.shape, Y_val.shape, Y_test.shape)
#create model
model = Sequential()
#get number of columns in training data
n_cols = X_train.shape[1]
#add model layers
model.add(Dense(32, activation='softmax', input_shape=(n_cols,)))
model.add(Dense(32, activation='relu'))
model.add(Dense(32, activation='softmax'))
model.add(Dense(1))
#compile model using rmsse as a measure of model performance
model.compile(optimizer='Adagrad', loss= "mean_absolute_error", metrics = ['accuracy'])
#set early stopping monitor so the model stops training when it won't improve anymore early_stopping_monitor = EarlyStopping(patience=3)
early_stopping_monitor = EarlyStopping(patience=20)
#train model
model.fit(X_train, Y_train,batch_size=32, epochs=10, validation_data=(X_val, Y_val))
Here is what I got
The plots are also pretty strange:
Accuracy
Loss
Two mistakes:
Accuracy is meaningless in regression settings, such as yours here (it is meaningful only for classification ones); see What function defines accuracy in Keras when the loss is mean squared error (MSE)? (the argument is identical when MAE loss is used, like here). Your performance measure here is the same with your loss (i.e. MAE).
We never use softmax activations in anything but the final layer of a classification model; replace both softmax activation functions used in your model with relu (keep the last layer as is, as no activation means linear, which is indeed the correct one for regression).
Beginner to Deep learning..
I'm trying to identify the slum using satellite images(google map) for Pune city. So, in training dataset i have provided about 100 images of slum and 100 images of other area. But my model is not able to classify input image properly even though accuracy rate is high.
I think this might be because of dimensions of image.
I'm resizing all images to 128*128 pixel.
Kernal size is 3*3.
Link to the map:
https://www.google.co.in/maps/#18.5129661,73.822531,286m/data=!3m1!1e3?hl=en
Following is the code
import os,cv2
import glob
import numpy as np
from keras.utils import plot_model
from keras.utils.np_utils import to_categorical
from sklearn.utils import shuffle
from sklearn.model_selection import train_test_split
from keras.models import Model
from keras.layers import Input, Convolution2D, MaxPooling2D, Flatten, Dense, Dropout
PATH = os.getcwd()
data_path = PATH + '/dataset/*'
files = glob.glob(data_path)
X = []
for myFiles in files:
image = cv2.imread(myFiles)
image_resize = cv2.resize(image, (256, 256))
X.append(image_resize)
image_data = np.array(X)
image_data = image_data.astype('float32')
image_data /= 255
print("Image_data shape ", image_data.shape)
no_of_classes = 2
no_of_samples = image_data.shape[0]
label = np.ones(no_of_samples, dtype='int64')
label[0:86] = 0 #Slum
label[87:] = 1 #noSlum
Y = to_categorical(label, no_of_classes)
#shuffle dataset
x,y = shuffle(image_data , Y, random_state = 2)
x_train, x_test, y_train, y_test = train_test_split(x, y, test_size=0.2, random_state = 2)
#print(x_train)
#print(y_train)
input_shape = image_data[0].shape
input = Input(input_shape)
conv_1 = Convolution2D(32,(3,3), padding='same', activation='relu')(input)
conv_2 = Convolution2D(32,(3,3), padding = 'same', activation = 'relu')(conv_1)
pool_1 = MaxPooling2D(pool_size = (2,2))(conv_2)
drop_1 = Dropout(0.5)(pool_1)
conv_3 = Convolution2D(64,(3,3), padding='same', activation='relu')(drop_1)
conv_4 = Convolution2D(64,(3,3), padding='same', activation = 'relu')(conv_3)
pool_2 = MaxPooling2D(pool_size = (2,2))(conv_4)
drop_2 = Dropout(0.5)(pool_2)
flat_1 = Flatten()(drop_2)
hidden = Dense(64,activation='relu')(flat_1)
drop_3 = Dropout(0.5)(hidden)
out = Dense(no_of_classes,activation = 'softmax')(drop_3)
model = Model(inputs = input, outputs = out)
model.compile(loss = 'categorical_crossentropy', optimizer = 'rmsprop', metrics= ['accuracy'])
model.fit(x_train,y_train,batch_size=10,nb_epoch=20,verbose =1, validation_data=(x_test,y_test))
model.save('model.h5')
score = model.evaluate(x_test,y_test,verbose=1)
print('Test Loss: ',score[0])
print('Test Accuracy: ',score[1])
test_image = x_test[0:1]
print(test_image.shape)
print (model.predict(test_image))
Usually, the behavior you've described above resembles to the inability of NN to identify small objects on input images. Just imagine you give an image of 128*128 with rough noise where nothing is seen - you want NN to correctly classify objects?
What to do?
1) Try to manually convert some input image from your dataset to 128*128 size and see on what data you truly train your NN. So, it'll give you more insight --> maybe you need to have better image's dimension size
2) Add more Conv layers with more neurons that will give you ability to detect small and more sophisticated objects by adding more non-linearity to your output function. Google such great Neural Network structures as ResNet.
3) Add more training data, 100 images isn't enough to have an appropriate result
4) Add data augmentation technique as well ( Rotations seem so strong in your case )
And don't give up :) Eventually, you'll solve it out. Good Luck
So this question is about GANs.
I am trying to do a trivial example for my own proof of concept; namely, generate images of hand written digits (MNIST). While most will approach this via deep convolutional gans (dgGANs), I am just trying to achieve this via the 1D array (i.e. instead of 28x28 gray-scale pixel values, a 28*28 1d array).
This git repo features a "vanilla" gans which treats the MNIST dataset as a 1d array of 784 values. Their output values look pretty acceptable so I wanted to do something similar.
Import statements
from __future__ import print_function
import matplotlib as mpl
from matplotlib import pyplot as plt
import mxnet as mx
from mxnet import nd, gluon, autograd
from mxnet.gluon import nn, utils
import numpy as np
import os
from math import floor
from random import random
import time
from datetime import datetime
import logging
ctx = mx.gpu()
np.random.seed(3)
Hyper parameters
batch_size = 100
epochs = 100
generator_learning_rate = 0.001
discriminator_learning_rate = 0.001
beta1 = 0.5
latent_z_size = 100
Load data
mnist = mx.test_utils.get_mnist()
# convert imgs to arrays
flattened_training_data = mnist["test_data"].reshape(10000, 28*28)
define models
G = nn.Sequential()
with G.name_scope():
G.add(nn.Dense(300, activation="relu"))
G.add(nn.Dense(28 * 28, activation="tanh"))
D = nn.Sequential()
with D.name_scope():
D.add(nn.Dense(128, activation="relu"))
D.add(nn.Dense(64, activation="relu"))
D.add(nn.Dense(32, activation="relu"))
D.add(nn.Dense(2, activation="tanh"))
loss = gluon.loss.SoftmaxCrossEntropyLoss()
init stuff
G.initialize(mx.init.Normal(0.02), ctx=ctx)
D.initialize(mx.init.Normal(0.02), ctx=ctx)
trainer_G = gluon.Trainer(G.collect_params(), 'adam', {"learning_rate": generator_learning_rate, "beta1": beta1})
trainer_D = gluon.Trainer(D.collect_params(), 'adam', {"learning_rate": discriminator_learning_rate, "beta1": beta1})
metric = mx.metric.Accuracy()
dynamic plot (for juptyer notebook)
import matplotlib.pyplot as plt
import time
def dynamic_line_plt(ax, y_data, colors=['r', 'b', 'g'], labels=['Line1', 'Line2', 'Line3']):
x_data = []
y_max = 0
y_min = 0
x_min = 0
x_max = 0
for y in y_data:
x_data.append(list(range(len(y))))
if max(y) > y_max:
y_max = max(y)
if min(y) < y_min:
y_min = min(y)
if len(y) > x_max:
x_max = len(y)
ax.set_ylim(y_min, y_max)
ax.set_xlim(x_min, x_max)
if ax.lines:
for i, line in enumerate(ax.lines):
line.set_xdata(x_data[i])
line.set_ydata(y_data[i])
else:
for i in range(len(y_data)):
l = ax.plot(x_data[i], y_data[i], colors[i], label=labels[i])
ax.legend()
fig.canvas.draw()
train
stamp = datetime.now().strftime('%Y_%m_%d-%H_%M')
logging.basicConfig(level=logging.DEBUG)
# arrays to store data for plotting
loss_D = nd.array([0], ctx=ctx)
loss_G = nd.array([0], ctx=ctx)
acc_d = nd.array([0], ctx=ctx)
labels = ['Discriminator Loss', 'Generator Loss', 'Discriminator Acc.']
%matplotlib notebook
fig, ax = plt.subplots(1, 1)
ax.set_xlabel('Time')
ax.set_ylabel('Loss')
dynamic_line_plt(ax, [loss_D.asnumpy(), loss_G.asnumpy(), acc_d.asnumpy()], labels=labels)
for epoch in range(epochs):
tic = time.time()
data_iter.reset()
for i, batch in enumerate(data_iter):
####################################
# Update Disriminator: maximize log(D(x)) + log(1-D(G(z)))
####################################
# extract batch of real data
data = batch.data[0].as_in_context(ctx)
# add noise
# Produce our noisey input to the generator
latent_z = mx.nd.random_normal(0,1,shape=(batch_size, latent_z_size), ctx=ctx)
# soft and noisy labels
# real_label = mx.nd.ones((batch_size, ), ctx=ctx) * nd.random_uniform(.7, 1.2, shape=(1)).asscalar()
# fake_label = mx.nd.ones((batch_size, ), ctx=ctx) * nd.random_uniform(0, .3, shape=(1)).asscalar()
# real_label = nd.random_uniform(.7, 1.2, shape=(batch_size), ctx=ctx)
# fake_label = nd.random_uniform(0, .3, shape=(batch_size), ctx=ctx)
real_label = mx.nd.ones((batch_size, ), ctx=ctx)
fake_label = mx.nd.zeros((batch_size, ), ctx=ctx)
with autograd.record():
# train with real data
real_output = D(data)
errD_real = loss(real_output, real_label)
# train with fake data
fake = G(latent_z)
fake_output = D(fake.detach())
errD_fake = loss(fake_output, fake_label)
errD = errD_real + errD_fake
errD.backward()
trainer_D.step(batch_size)
metric.update([real_label, ], [real_output,])
metric.update([fake_label, ], [fake_output,])
####################################
# Update Generator: maximize log(D(G(z)))
####################################
with autograd.record():
output = D(fake)
errG = loss(output, real_label)
errG.backward()
trainer_G.step(batch_size)
####
# Plot Loss
####
# append new data to arrays
loss_D = nd.concat(loss_D, nd.mean(errD), dim=0)
loss_G = nd.concat(loss_G, nd.mean(errG), dim=0)
name, acc = metric.get()
acc_d = nd.concat(acc_d, nd.array([acc], ctx=ctx), dim=0)
# plot array
dynamic_line_plt(ax, [loss_D.asnumpy(), loss_G.asnumpy(), acc_d.asnumpy()], labels=labels)
name, acc = metric.get()
metric.reset()
logging.info('Binary training acc at epoch %d: %s=%f' % (epoch, name, acc))
logging.info('time: %f' % (time.time() - tic))
output
img = G(mx.nd.random_normal(0,1,shape=(100, latent_z_size), ctx=ctx))[0].reshape((28, 28))
plt.imshow(img.asnumpy(),cmap='gray')
plt.show()
Now this doesn't get nearly as good as the repo's example from above. Although fairly similar.
Thus I was wondering if you could take a look and figure out why:
the colors are inverted
why the results are sub par
I have been fiddling around with this trying a lot of various things to improve the results (I will list this in a second), but for the MNIST dataset this really shouldn't be needed.
Things I have tried (and I have also tried a host of combinations):
increasing the generator network
increasing the discriminator network
using soft labeling
using noisy labeling
batch norm after every layer in the generator
batch norm of the data
normalizing all values between -1 and 1
leaky relus in the generator
drop out layers in the generator
increased learning rate of discriminator compared to generator
decreased learning rate of i compared to generator
Please let me know if you have any ideas.
1) If you look into original dataset:
training_set = mnist["train_data"].reshape(60000, 28, 28)
plt.imshow(training_set[10,:,:], cmap='gray')
you will notice that the digits are white on a black background. So, technically speaking, your results are not inversed - they match the pattern of original images you used as a real data.
If you want to invert colors for visualization purposes, you can easily do that by changing the pallete to reversed one by adding '_r' (it works for all color palletes):
plt.imshow(img.asnumpy(), cmap='gray_r')
You also can play with ranges of colors by changing vmin and vmax parameters. They control how big the difference between colors should be. By default it is calculated automatically based on provided set.
2) "Why the results are sub par" - I think this is exactly the reason why the community started to use dcGANs. To me the results in the git repo you provided are quite noisy. Surely, they are different from what you receive, and you can achieve the same quality just by changing your activation functions from tanh to sigmoid as in the example on github:
G = nn.Sequential()
with G.name_scope():
G.add(nn.Dense(300, activation="relu"))
G.add(nn.Dense(28 * 28, activation="sigmoid"))
D = nn.Sequential()
with D.name_scope():
D.add(nn.Dense(128, activation="relu"))
D.add(nn.Dense(64, activation="relu"))
D.add(nn.Dense(32, activation="relu"))
D.add(nn.Dense(2, activation="sigmoid"))
Sigmoid never goes below zero and it works better in this scenario. Here is a sample picture I get if I train updated model for 30 epochs (the rest of the hyperparameters are same).
If you decide to explore dcGAN to get even better results, take a look here - https://mxnet.incubator.apache.org/tutorials/unsupervised_learning/gan.html It is a well explained tutorial on how to build dcGAN with Mxnet and Gluon. By using dcGAN you will get way better results than that.
Consider this minimal runnable example:
from keras.models import Sequential
from keras.layers import Dense
from keras.layers import LSTM
import numpy as np
import matplotlib.pyplot as plt
max = 30
step = 0.5
n_steps = int(30/0.5)
x = np.arange(0,max,step)
x = np.cos(x)*(max-x)/max
y = np.roll(x,-1)
y[-1] = x[-1]
shape = (n_steps,1,1)
batch_shape = (1,1,1)
x = x.reshape(shape)
y = y.reshape(shape)
model = Sequential()
model.add(LSTM(50, return_sequences=True, stateful=True, batch_input_shape=batch_shape))
model.add(LSTM(50, return_sequences=True, stateful=True))
model.add(Dense(1))
model.compile(loss='mse', optimizer='rmsprop')
for i in range(1000):
model.reset_states()
model.fit(x,y,nb_epoch=1, batch_size=1)
p = model.predict(x, batch_size=1)
plt.clf()
plt.axis([-1,31, -1.1, 1.1])
plt.plot(x[:, 0, 0], '*')
plt.plot(y[:,0,0],'o')
plt.plot(p[:,0,0],'.')
plt.draw()
plt.pause(0.001)
As stated in the keras API https://keras.io/layers/recurrent/
the last state for each sample at index i in a batch will be used as
initial state for the sample of index i in the following batch
So I'm using batch_size = 1 and I'm trying to predict the next value in the decaying cos-function for each timestep. The prediction, or the red dots in the picture below should go into the green circles for the script to predict it correctly, however it doesn't converge... Have any idea to make it learn?
The problem lied in a calling model.fit for each epoch separately. In this case optimizer parameters are reset what was harmful for a training process. Other thing is calling reset_states also before prediction - as if it wasn't called - the states from fit are starting states for prediction what also might be harmful. The final code is following:
for epoch in range(1000):
model.reset_states()
tot_loss = 0
for batch in range(n_steps):
batch_loss = model.train_on_batch(x[batch:batch+1], y[batch:batch+1])
tot_loss+=batch_loss
print "Loss: " + str(tot_loss/float(n_steps))
model.reset_states()
p = model.predict(x, batch_size=1)
What is an example of how to use a TensorFlow TFRecord with a Keras Model and tf.session.run() while keeping the dataset in tensors w/ queue runners?
Below is a snippet that works but it needs the following improvements:
Use the Model API
specify an Input()
Load a dataset from a TFRecord
Run through a dataset in parallel (such as with a queuerunner)
Here is the snippet, there are several TODO lines indicating what is needed:
from keras.models import Model
import tensorflow as tf
from keras import backend as K
from keras.layers import Dense, Input
from keras.objectives import categorical_crossentropy
from tensorflow.examples.tutorials.mnist import input_data
sess = tf.Session()
K.set_session(sess)
# Can this be done more efficiently than placeholders w/ TFRecords?
img = tf.placeholder(tf.float32, shape=(None, 784))
labels = tf.placeholder(tf.float32, shape=(None, 10))
# TODO: Use Input()
x = Dense(128, activation='relu')(img)
x = Dense(128, activation='relu')(x)
preds = Dense(10, activation='softmax')(x)
# TODO: Construct model = Model(input=inputs, output=preds)
loss = tf.reduce_mean(categorical_crossentropy(labels, preds))
# TODO: handle TFRecord data, is it the same?
mnist_data = input_data.read_data_sets('MNIST_data', one_hot=True)
train_step = tf.train.GradientDescentOptimizer(0.5).minimize(loss)
sess.run(tf.global_variables_initializer())
# TODO remove default, add queuerunner
with sess.as_default():
for i in range(1000):
batch = mnist_data.train.next_batch(50)
train_step.run(feed_dict={img: batch[0],
labels: batch[1]})
print(loss.eval(feed_dict={img: mnist_data.test.images,
labels: mnist_data.test.labels}))
Why is this question relevant?
For high performance training without going back to python
no TFRecord to numpy to tensor conversions
Keras will soon be part of tensorflow
Demonstrate how Keras Model() classes can accept tensors for input data correctly.
Here is some starter information for a semantic segmentation problem example:
example unet Keras model unet.py, happens to be for semantic segmentation.
Keras + Tensorflow Blog Post
An attempt at running the unet model a tf session with TFRecords and a Keras model (not working)
Code to create the TFRecords: tf_records.py
An attempt at running the unet model a tf session with TFRecords and a Keras model is in densenet_fcn.py (not working)
I don't use tfrecord dataset format so won't argue on the pros and cons, but I got interested in extending Keras to support the same.
github.com/indraforyou/keras_tfrecord is the repository. Will briefly explain the main changes.
Dataset creation and loading
data_to_tfrecord and read_and_decode here takes care of creating tfrecord dataset and loading the same. Special care must be to implement the read_and_decode otherwise you will face cryptic errors during training.
Initialization and Keras model
Now both tf.train.shuffle_batch and Keras Input layer returns tensor. But the one returned by tf.train.shuffle_batch don't have metadata needed by Keras internally. As it turns out, any tensor can be easily turned into a tensor with keras metadata by calling Input layer with tensor param.
So this takes care of initialization:
x_train_, y_train_ = ktfr.read_and_decode('train.mnist.tfrecord', one_hot=True, n_class=nb_classes, is_train=True)
x_train_batch, y_train_batch = K.tf.train.shuffle_batch([x_train_, y_train_],
batch_size=batch_size,
capacity=2000,
min_after_dequeue=1000,
num_threads=32) # set the number of threads here
x_train_inp = Input(tensor=x_train_batch)
Now with x_train_inp any keras model can be developed.
Training (simple)
Lets say train_out is the output tensor of your keras model. You can easily write a custom training loop on the lines of:
loss = tf.reduce_mean(categorical_crossentropy(y_train_batch, train_out))
train_op = tf.train.GradientDescentOptimizer(0.01).minimize(loss)
# sess.run(tf.global_variables_initializer())
sess.run(tf.initialize_all_variables())
with sess.as_default():
coord = tf.train.Coordinator()
threads = tf.train.start_queue_runners(sess=sess, coord=coord)
try:
step = 0
while not coord.should_stop():
start_time = time.time()
_, loss_value = sess.run([train_op, loss], feed_dict={K.learning_phase(): 0})
duration = time.time() - start_time
if step % 100 == 0:
print('Step %d: loss = %.2f (%.3f sec)' % (step, loss_value,
duration))
step += 1
except tf.errors.OutOfRangeError:
print('Done training for %d epochs, %d steps.' % (FLAGS.num_epochs, step))
finally:
coord.request_stop()
coord.join(threads)
sess.close()
Training (keras style)
One of the features of keras that makes it so lucrative is its generalized training mechanism with the callback functions.
But to support tfrecords type training there are several changes that are need in the fit function
running the queue threads
no feeding in batch data through feed_dict
supporting validation becomes tricky as the validation data will also be coming in through another tensor an different model needs to be internally created with shared upper layers and validation tensor fed in by other tfrecord reader.
But all this can be easily supported by another flag parameter. What makes things messing are the keras features sample_weight and class_weight they are used to weigh each sample and weigh each class. For this in compile() keras creates placeholders (here) and placeholders are also implicitly created for the targets (here) which is not needed in our case the labels are already fed in by tfrecord readers. These placeholders needs to be fed in during session run which is unnecessary in our cae.
So taking into account these changes, compile_tfrecord(here) and fit_tfrecord(here) are the extension of compile and fit and shares say 95% of the code.
They can be used in the following way:
import keras_tfrecord as ktfr
train_model = Model(input=x_train_inp, output=train_out)
ktfr.compile_tfrecord(train_model, optimizer='rmsprop', loss='categorical_crossentropy', out_tensor_lst=[y_train_batch], metrics=['accuracy'])
train_model.summary()
ktfr.fit_tfrecord(train_model, X_train.shape[0], batch_size, nb_epoch=3)
train_model.save_weights('saved_wt.h5')
You are welcome to improve on the code and pull requests.
Update 2018-08-29 this is now directly supported in keras, see the following example:
https://github.com/keras-team/keras/blob/master/examples/mnist_tfrecord.py
Original Answer:
TFRecords are supported by using an external loss. Here are the key lines constructing an external loss:
# tf yield ops that supply dataset images and labels
x_train_batch, y_train_batch = read_and_decode_recordinput(...)
# create a basic cnn
x_train_input = Input(tensor=x_train_batch)
x_train_out = cnn_layers(x_train_input)
model = Model(inputs=x_train_input, outputs=x_train_out)
loss = keras.losses.categorical_crossentropy(y_train_batch, x_train_out)
model.add_loss(loss)
model.compile(optimizer='rmsprop', loss=None)
Here is an example for Keras 2. It works after applying the small patch #7060:
'''MNIST dataset with TensorFlow TFRecords.
Gets to 99.25% test accuracy after 12 epochs
(there is still a lot of margin for parameter tuning).
'''
import os
import copy
import time
import numpy as np
import tensorflow as tf
from tensorflow.python.ops import data_flow_ops
from keras import backend as K
from keras.models import Model
from keras.layers import Dense
from keras.layers import Dropout
from keras.layers import Flatten
from keras.layers import Input
from keras.layers import Conv2D
from keras.layers import MaxPooling2D
from keras.callbacks import EarlyStopping
from keras.callbacks import TensorBoard
from keras.objectives import categorical_crossentropy
from keras.utils import np_utils
from keras.utils.generic_utils import Progbar
from keras import callbacks as cbks
from keras import optimizers, objectives
from keras import metrics as metrics_module
from keras.datasets import mnist
if K.backend() != 'tensorflow':
raise RuntimeError('This example can only run with the '
'TensorFlow backend for the time being, '
'because it requires TFRecords, which '
'are not supported on other platforms.')
def images_to_tfrecord(images, labels, filename):
def _int64_feature(value):
return tf.train.Feature(int64_list=tf.train.Int64List(value=[value]))
def _bytes_feature(value):
return tf.train.Feature(bytes_list=tf.train.BytesList(value=[value]))
""" Save data into TFRecord """
if not os.path.isfile(filename):
num_examples = images.shape[0]
rows = images.shape[1]
cols = images.shape[2]
depth = images.shape[3]
print('Writing', filename)
writer = tf.python_io.TFRecordWriter(filename)
for index in range(num_examples):
image_raw = images[index].tostring()
example = tf.train.Example(features=tf.train.Features(feature={
'height': _int64_feature(rows),
'width': _int64_feature(cols),
'depth': _int64_feature(depth),
'label': _int64_feature(int(labels[index])),
'image_raw': _bytes_feature(image_raw)}))
writer.write(example.SerializeToString())
writer.close()
else:
print('tfrecord %s already exists' % filename)
def read_and_decode_recordinput(tf_glob, one_hot=True, classes=None, is_train=None,
batch_shape=[1000, 28, 28, 1], parallelism=1):
""" Return tensor to read from TFRecord """
print 'Creating graph for loading %s TFRecords...' % tf_glob
with tf.variable_scope("TFRecords"):
record_input = data_flow_ops.RecordInput(
tf_glob, batch_size=batch_shape[0], parallelism=parallelism)
records_op = record_input.get_yield_op()
records_op = tf.split(records_op, batch_shape[0], 0)
records_op = [tf.reshape(record, []) for record in records_op]
progbar = Progbar(len(records_op))
images = []
labels = []
for i, serialized_example in enumerate(records_op):
progbar.update(i)
with tf.variable_scope("parse_images", reuse=True):
features = tf.parse_single_example(
serialized_example,
features={
'label': tf.FixedLenFeature([], tf.int64),
'image_raw': tf.FixedLenFeature([], tf.string),
})
img = tf.decode_raw(features['image_raw'], tf.uint8)
img.set_shape(batch_shape[1] * batch_shape[2])
img = tf.reshape(img, [1] + batch_shape[1:])
img = tf.cast(img, tf.float32) * (1. / 255) - 0.5
label = tf.cast(features['label'], tf.int32)
if one_hot and classes:
label = tf.one_hot(label, classes)
images.append(img)
labels.append(label)
images = tf.parallel_stack(images, 0)
labels = tf.parallel_stack(labels, 0)
images = tf.cast(images, tf.float32)
images = tf.reshape(images, shape=batch_shape)
# StagingArea will store tensors
# across multiple steps to
# speed up execution
images_shape = images.get_shape()
labels_shape = labels.get_shape()
copy_stage = data_flow_ops.StagingArea(
[tf.float32, tf.float32],
shapes=[images_shape, labels_shape])
copy_stage_op = copy_stage.put(
[images, labels])
staged_images, staged_labels = copy_stage.get()
return images, labels
def save_mnist_as_tfrecord():
(X_train, y_train), (X_test, y_test) = mnist.load_data()
X_train = X_train[..., np.newaxis]
X_test = X_test[..., np.newaxis]
images_to_tfrecord(images=X_train, labels=y_train, filename='train.mnist.tfrecord')
images_to_tfrecord(images=X_test, labels=y_test, filename='test.mnist.tfrecord')
def cnn_layers(x_train_input):
x = Conv2D(32, (3, 3), activation='relu', padding='valid')(x_train_input)
x = Conv2D(64, (3, 3), activation='relu')(x)
x = MaxPooling2D(pool_size=(2, 2))(x)
x = Dropout(0.25)(x)
x = Flatten()(x)
x = Dense(128, activation='relu')(x)
x = Dropout(0.5)(x)
x_train_out = Dense(classes,
activation='softmax',
name='x_train_out')(x)
return x_train_out
sess = tf.Session()
K.set_session(sess)
save_mnist_as_tfrecord()
batch_size = 100
batch_shape = [batch_size, 28, 28, 1]
epochs = 3000
classes = 10
parallelism = 10
x_train_batch, y_train_batch = read_and_decode_recordinput(
'train.mnist.tfrecord',
one_hot=True,
classes=classes,
is_train=True,
batch_shape=batch_shape,
parallelism=parallelism)
x_test_batch, y_test_batch = read_and_decode_recordinput(
'test.mnist.tfrecord',
one_hot=True,
classes=classes,
is_train=True,
batch_shape=batch_shape,
parallelism=parallelism)
x_batch_shape = x_train_batch.get_shape().as_list()
y_batch_shape = y_train_batch.get_shape().as_list()
x_train_input = Input(tensor=x_train_batch, batch_shape=x_batch_shape)
x_train_out = cnn_layers(x_train_input)
y_train_in_out = Input(tensor=y_train_batch, batch_shape=y_batch_shape, name='y_labels')
cce = categorical_crossentropy(y_train_batch, x_train_out)
train_model = Model(inputs=[x_train_input], outputs=[x_train_out])
train_model.add_loss(cce)
train_model.compile(optimizer='rmsprop',
loss=None,
metrics=['accuracy'])
train_model.summary()
tensorboard = TensorBoard()
# tensorboard disabled due to Keras bug
train_model.fit(batch_size=batch_size,
epochs=epochs) # callbacks=[tensorboard])
train_model.save_weights('saved_wt.h5')
K.clear_session()
# Second Session, pure Keras
(X_train, y_train), (X_test, y_test) = mnist.load_data()
X_train = X_train[..., np.newaxis]
X_test = X_test[..., np.newaxis]
x_test_inp = Input(batch_shape=(None,) + (X_test.shape[1:]))
test_out = cnn_layers(x_test_inp)
test_model = Model(inputs=x_test_inp, outputs=test_out)
test_model.load_weights('saved_wt.h5')
test_model.compile(optimizer='rmsprop', loss='categorical_crossentropy', metrics=['accuracy'])
test_model.summary()
loss, acc = test_model.evaluate(X_test, np_utils.to_categorical(y_test), classes)
print('\nTest accuracy: {0}'.format(acc))
I've also been working to improve the support for TFRecords in the following issue and pull request:
#6928 Yield Op support: High Performance Large Datasets via TFRecords, and RecordInput
#7102 Keras Input Tensor API Design Proposal
Finally, it is possible to use tf.contrib.learn.Experiment to train Keras models in TensorFlow.