I'm currently experimenting with the new high level tf.contrib.learn APIs in Tensorflow by porting over features from the inception-v3 retrain script from this tutorial:
May I know how do I replicate the minibatch sampling every iteration for both validation and training inputs as seen in the original retrain.py?
Currently I tried using tf.train.shuffle_batch in the input_fn function but not I'm not certain that it works.
Some code snippet for clarity
def train_input_fn():
# Get a batch of input bottleneck values, either calculated fresh every
# time with distortions applied, or from the cache stored on disk.
if do_distort_images:
(train_bottleneck_outputs, train_ground_truths) = get_random_distorted_bottlenecks(
sess, image_lists, -1, 'training',
FLAGS.image_dir, distorted_jpeg_data_tensor,
distorted_image_tensor, resized_image_tensor, bottleneck_tensor)
else:
(train_bottleneck_outputs, train_ground_truths, _) = get_random_cached_bottlenecks(
sess, image_lists, -1, 'training',
FLAGS.bottleneck_dir, FLAGS.image_dir, jpeg_data_tensor,
bottleneck_tensor)
return tf.train.shuffle_batch([tf.constant(train_bottleneck_outputs),
tf.constant(train_ground_truths)],
batch_size=FLAGS.train_batch_size, capacity=1100,
min_after_dequeue=1000, enqueue_many=True, num_threads=2)
Related
I’m dealing with CIFAR10 and I use torchvision.datasets to create it. I’m in need of GPU to accelerate the calculation but I can’t find a way to put the whole dataset into GPU at one time. My model need to use mini-batches and it is really time-consuming to deal with each batch separately.
I've tried to put each mini-batch into GPU separately but it seems really time-consuming.
TL;DR
You won't save time by moving the entire dataset at once.
I don't think you'd necessarily want to do that even if you have the GPU memory to handle the entire dataset (of course, CIFAR10 is tiny by today's standards).
I tried various batch sizes and timed the transfer to GPU as follows:
num_workers = 1 # Set this as needed
def time_gpu_cast(batch_size=1):
start_time = time()
for x, y in DataLoader(dataset, batch_size, num_workers=num_workers):
x.cuda(); y.cuda()
return time() - start_time
# Try various batch sizes
cast_times = [(2 ** bs, time_gpu_cast(2 ** bs)) for bs in range(15)]
# Try the entire dataset like you want to do
cast_times.append((len(dataset), time_gpu_cast(len(dataset))))
plot(*zip(*cast_times)) # Plot the time taken
For num_workers = 1, this is what I got:
And if we try parallel loading (num_workers = 8), it becomes even clearer:
I've got an answer and I'm gonna try it later. It seems promising.
You can write a dataset class where in the init function, you red the entire dataset and apply all the transformations you need, and convert them to tensor format. Then, send this tensor to GPU (assuming there is enough memory). Then, in the getitem function you can simply use the index to retrieve the elements of that tensor which is already on GPU.
I believe that deeplearning4j and R with exactly the same parameters should perform the same, comparable MSE. But I am not sure how to achieve that.
I have a csv file with the following format, which contains 46 variables and 2 outputs. Totally there are 1,0000 samples. All the data is normalized and the model is for regression analysis.
S1 | S2 | ... | S46 | X | Y
In R, I use neuralnet package, and the code is:
rn <- colnames(traindata)
f <- as.formula(paste("X + Y ~", paste(rn[1:(length(rn)-2)], collapse="+")))
nn <- neuralnet(f,
rep=1,
data=traindata,
hidden=c(10),
linear.output=T,
threshold = 0.5)
which is quite straightforward.
As I want to integrate the algorithm into Java project, so I consider dl4j to train the model. The trainset is exactly the same as that in R code. The test set is randomly selected. THe dl4j code is:
MultiLayerConfiguration conf = new NeuralNetConfiguration.Builder()
.seed(rngSeed) //include a random seed for reproducibility
// use stochastic gradient descent as an optimization algorithm
.optimizationAlgo(OptimizationAlgorithm.STOCHASTIC_GRADIENT_DESCENT)
.iterations(100)
.learningRate(0.0001) //specify the learning rate
.updater(Updater.NESTEROVS).momentum(0.9) //specify the rate of change of the learning rate.
.regularization(true).l2(0.0001)
.list()
.layer(0, new DenseLayer.Builder() //create the first, input layer with xavier initialization
.nIn(46)
.nOut(10)
.activation(Activation.TANH)
.weightInit(WeightInit.XAVIER)
.build())
.layer(1, new OutputLayer.Builder(LossFunctions.LossFunction.MSE) //create hidden layer
.nIn(10)
.nOut(outputNum)
.activation(Activation.IDENTITY)
.build())
.pretrain(false).backprop(true) //use backpropagation to adjust weights
.build();
The number of epoch is 10 and batchsize is 128.
Using the test set, the performance of R is
and the performance of dl4j is the following, I think it does not work out its full potential.
The mornitor of dl4j is
As there is much more parameters in dl4j such as updater, regulization and weightInit. So I think some of the parameters are not properly set. BTW, why there are periodic thorns in the mornitor graph.
Can any one help?
Most neural net training happens in minibatches. Deeplearning4j assumes you aren't doing toy problems by default (all data in memory < 10 examples etc)
The neural net config has a function called minibatch you should look for.
Set minibatch to false on the configuration and you should get the same results.
If you're wondering why this happens, it's because minibatch learning is different from doing everything in memory. Minibatch learning automatically divides the gradient by the minibatch size. When you do everything in memory you don't want that.
Take note of this when you're running other experiments.
For more on this see:
https://deeplearning4j.org/toyproblems
I am playing around with Tensorflow and there is one thing I can't seem to understand: the role of the fetches argument for the Session.run().
From the doc:
To fetch the outputs of operations, execute the graph with a run()
call on the Session object and pass in the tensors to retrieve.
What I understood is that the parameter is used to read certain values from the from the run step. I see, though, that different fetches bring to different results
....´
loss = tf.reduce_mean(
tf.nn.softmax_cross_entropy_with_logits(preactivations_output, tf_train_labels))
optimizer = tf.train.GradientDescentOptimizer(0.5).minimize(loss)
...
train_prediction = tf.nn.softmax(preactivations_output)
...
# This produces a model with 89% accuracy
_, l, predictions = session.run(
[optimizer, loss, train_prediction], feed_dict=feed_dict)
# This produces a model with 10% accuracy
l, predictions = session.run(
[loss, train_prediction], feed_dict=feed_dict)
So now I wonder what else the fetches parameter is used for, given that it seems to influence the way the training is run. I didn't really understand it from the documentation...
I am involved with an application that needs to estimate the state of a certain system in real time by measuring a set of (non-linearly) dependent parameters. Up until now the application was using an extended Kalman filter, but it was found to be underperforming in certain circumstances, which is likely caused by the fact that the differences between the real system and its model used in the filter are too significant to be modeled as white noise. We cannot use a more precise model for a number of unrelated reasons.
We decided to try recurrent neural networks for the task. Since my experience with neural networks is quite limited, before tackling the real task itself, I decided to practice with a hand crafted problem first. That problem, however, I could not solve, so I'm asking for help here.
Here's what I did: I generated some sine waveforms of varying phase, frequency, amplitude, and offset. Then I distorted the waveforms with some white noise, and (unsuccessfully) attempted to train an LSTM network to recover my waveforms from the noisy signal. I expected that the network will eventually learn to fit a sine waveform into the noisy data set.
Here's the source (slightly abridged, but it should work):
#!/usr/bin/env python3
import time
import numpy as np
from keras.models import Sequential
from keras.layers import Dense, LSTM
from keras.layers.wrappers import TimeDistributed
from keras.objectives import mean_absolute_error, cosine_proximity
POINTS_PER_WF = int(1e4)
X_SPACE = np.linspace(0, 100, POINTS_PER_WF)
def make_waveform_with_noise():
def add_noise(vec):
stdev = float(np.random.uniform(0.01, 0.2))
return vec + np.random.normal(0, stdev, size=len(vec))
f = np.random.choice((np.sin, np.cos))
wf = f(X_SPACE * np.random.normal(scale=5)) *\
np.random.normal(scale=5) + np.random.normal(scale=50)
return wf, add_noise(wf)
RESCALING = 1e-3
BATCH_SHAPE = (1, POINTS_PER_WF, 1)
model = Sequential([
TimeDistributed(Dense(5, activation='tanh'), batch_input_shape=BATCH_SHAPE),
LSTM(20, activation='tanh', inner_activation='sigmoid', return_sequences=True),
LSTM(20, activation='tanh', inner_activation='sigmoid', return_sequences=True),
TimeDistributed(Dense(1, activation='tanh'))
])
def compute_loss(y_true, y_pred):
skip_first = POINTS_PER_WF // 2
y_true = y_true[:, skip_first:, :] * RESCALING
y_pred = y_pred[:, skip_first:, :] * RESCALING
me = mean_absolute_error(y_true, y_pred)
cp = cosine_proximity(y_true, y_pred)
return me + cp
model.summary()
model.compile(optimizer='adam', loss=compute_loss,
metrics=['mae', 'cosine_proximity'])
NUM_ITERATIONS = 30000
for iteration in range(NUM_ITERATIONS):
wf, noisy_wf = make_waveform_with_noise()
y = wf.reshape(BATCH_SHAPE) * RESCALING
x = noisy_wf.reshape(BATCH_SHAPE) * RESCALING
info = model.train_on_batch(x, y)
model.save_weights('final.hdf5')
The first dense layer is actually useless, the reason I added it is because I wanted to make sure I can successfully combine LSTM and time distributed dense layers, since my real application will likely need that setup.
The error function was modified a number of times. Initially I was using plain mean squared error, but the training process was extremely slow, and it was mostly converging to simply copying the input noisy signal into the output. The cosine proximity metric I added later essentially defines the degree of similarity between the shapes of the functions; it seemed to speed up the learning quite a bit. Also note that I'm applying the loss function only to the last half of the dataset; the motivation for that is that I expected that the network will need to see a few periods of the signal in order to be able to correctly identify the parameters of the waveform. However, I found that this modification has no visible effect on the performance of the network.
The latest modification of the script uses Adam optimizer, I also experimented with RMSProp with varying learning rate and decay settings, but I found no noticeable difference in behavior of the network.
I am using Theano 0.9 (dev) backend configured to use 64 bit floating point, in order to prevent possible issues with numerical stability. The epsilon value is set accordingly to 1e-14.
This is what the output looks like after 15k..30k training steps (performance stops improving starting from about 15k steps) (the first plot is zoomed in for the sake of clarity):
Plot legend:
blue (0) - noisy signal, input of the RNN
green (1) - recovered signal, output of the RNN
red (2) - ground truth
My question is: what am I doing wrong?
The general idea I am trying to realize is a seq2seq-model (taken from the translate.py-example in the models, based on the seq2seq-class). This trains well.
Furthermore I am using the hidden state of the rnn after all the encoding is done, right before decoding starts (I call it the “hidden state at end of encoding”). I use this hidden state at end of encoding to feed it into a further sub-graph which I call “prices” (see below). The training gradients of this sub-graph backprop not only through this additional sub-graph, but also back into the encoder-part of the rnn (which is what I want and need).
The plan is to add more such sub-graph to the hidden state at end of encoding, as I want to analyze the input phrases in a variety of ways.
Now during training when I evaluate and train both sub-graphs (encoder+prices AND encoder+decoder) at the same time, the net does NOT converge. However, if I train by executing the training in the following way (pseudo-code):
if global_step % 10 == 0:
execute-the-price-training_code
else:
execute-the-decoder-training_code
So I am not training both sub-graphs simultaneously. Now it does converge, but the encoder+decoder-part converges MUCH slower than if I ONLY train this part and never train the prices-sub-graph.
My question is: I should be able to train both sub-graphs simultaneously. But probably I have to rescale the gradients flowing back into the hidden state at end of encoding. Here we get the gradients from the prices sub-graph AND from the decoder-sub-graph. How should this rescaling be done. I didnt find any papers describing such an undertaking, but maybe I am searching with the wrong keywords.
Here is the training-part of the code:
This is the (almost original) training-op-preparation:
if not forward_only:
self.gradient_norms = []
self.updates = []
opt = tf.train.AdadeltaOptimizer(self.learning_rate)
for bucket_id in xrange(len(buckets)):
tf.scalar_summary("seq2seq loss", self.losses[bucket_id])
gradients = tf.gradients(self.losses[bucket_id], var_list_seq2seq)
clipped_gradients, norm = tf.clip_by_global_norm(gradients, max_gradient_norm)
self.gradient_norms.append(norm)
self.updates.append(opt.apply_gradients(zip(clipped_gradients, var_list_seq2seq), global_step=self.global_step))
Now, additionally, I am running a second sub-graph that takes the hidden state at end of encoding as input:
with tf.name_scope('prices') as scope:
#First layer
W_price_first_layer = tf.Variable(tf.random_normal([num_layers*size, self.prices_hidden_layer_size], stddev=0.35), name="W_price_first_layer")
B_price_first_layer = tf.Variable(tf.zeros([self.prices_hidden_layer_size]), name="B_price_first_layer")
self.output_price_first_layer = tf.add(tf.matmul(self.hidden_state, W_price_first_layer), B_price_first_layer)
self.activation_price_first_layer = tf.nn.sigmoid(self.output_price_first_layer)
#self.activation_price_first_layer = tf.nn.Relu(self.output_price_first_layer)
#Second layer to softmax (price ranges)
W_price = tf.Variable(tf.random_normal([self.prices_hidden_layer_size, self.prices_bit_size], stddev=0.35), name="W_price")
W_price_t = tf.transpose(W_price)
B_price = tf.Variable(tf.zeros([self.prices_bit_size]), name="B_price")
self.output_price_second_layer = tf.add(tf.matmul(self.activation_price_first_layer, W_price),B_price)
self.price_prediction = tf.nn.softmax(self.output_price_second_layer)
self.label_price = tf.placeholder(tf.int32, shape=[self.batch_size], name="price_label")
#Remember the prices trainables
var_list_prices = tf.get_collection(tf.GraphKeys.TRAINABLE_VARIABLES, "prices")
var_list_all = tf.trainable_variables()
#Backprop
self.loss_price = tf.nn.sparse_softmax_cross_entropy_with_logits(self.output_price_second_layer, self.label_price)
self.loss_price_scalar = tf.reduce_mean(self.loss_price)
self.optimizer_price = tf.train.AdadeltaOptimizer(self.learning_rate_prices)
self.training_op_price = self.optimizer_price.minimize(self.loss_price, var_list=var_list_all)
Thx a bunch
I expect that running two optimizers simultaneously will lead to inconsistent gradient updates on the common variables, and this might be causing your training not to converge.
Instead, if you add the scalar loss from each sub-network to the "losses collection" (e.g. via tf.contrib.losses.add_loss() or tf.add_to_collection(tf.GraphKeys.LOSSES, ...), you can use tf.contrib.losses.get_total_loss() to get a single loss value that can be passed to a single standard TensorFlow tf.train.Optimizer subclass. TensorFlow will derive the appropriate back-prop computation for your split network.
The get_total_loss() method simply computes an unweighted sum of the values that have been added to the losses collection. I'm not familiar with the literature on how or if you should scale these values, but you can use any arbitrary (differentiable) TensorFlow expression to combine the losses and pass the result to a single optimizer.