I am working on the binary classification. I have created my network like:
Conv1, Relu1, Pool1 - Conv2, Relu2, Pool2 - Conv3, Relu3, Pool3 - Conv4, Relu4 - Conv5 Relu5 Dropot 0.5, FC, Dropout 0.5 - SoftmaxlossLayer
All conv layer is 3x3.
The default Weightdecay is 0.0005. And I am getting this result.
Training accuracy: 98%
Testing Accuracy: 88%
The same network is then used with Weightdecay 0.005
Anyone, Please help me to share why it is showing like that by changing the weight decay value?
Weight decay penalizes model complexity, so it's used to control model's variance against its bias. Clearly if you penalize complexity too much, the model will not learn anything useful, since it will be too simple.
For other methods of regularizing neural networks you can see notes for Hinton's Coursera course.
Related
I am using Conv-LSTM for training, and the input features have been proven to be effective in some papers, and I can use CNN+FC networks to extract features and classify them. I change the task to regression here, and I can also achieve model convergence with Conv+FC. Later, I tried to use Conv-LSTM for processing to consider the timing characteristics of the corresponding data. Specifically: return the output of the current moment based on multiple historical inputs and the input of the current moment. The Conv-LSTM code I used: https://github.com/ndrplz/ConvLSTM_pytorch. My Loss is L1-Loss and optimizer is Adam.
A loss curve is below:
Example loss value:
Epoch:1/500 AVG Training Loss:16.40108 AVG Valid Loss:22.40100
Best validation loss: 22.400997797648113
Saving best model for epoch 1
Epoch:2/500 AVG Training Loss:16.42522 AVG Valid Loss:22.40100
Epoch:3/500 AVG Training Loss:16.40599 AVG Valid Loss:22.40100
Epoch:4/500 AVG Training Loss:16.40175 AVG Valid Loss:22.40100
Epoch:5/500 AVG Training Loss:16.42198 AVG Valid Loss:22.40101
Epoch:6/500 AVG Training Loss:16.41907 AVG Valid Loss:22.40101
Epoch:7/500 AVG Training Loss:16.42531 AVG Valid Loss:22.40101
My attempt:
Adjust the data set to only a few samples, verify that it can be overfitted, and the network code should be fine.
Adjusting the learning rate, I tried 1e-3, 1e-4, 1e-5 and 1e-6, but the loss curve is still flat as before, and even the value of the loss curve has not changed much.
Replace the optimizer with SGD, and the training result is also the above problem.
Because my data is wireless data (I-Q), neither CV nor NLP input type, here are some questions to ask about deep learning training.
After some testing, I finally found that my initial learning rate was too small. According to my previous single-point data training, the learning rate of 1e-3 is large enough, so here is preconceived, and it is adjusted from 1e-3 to a small tune, but in fact, the learning rate of 1e-3 is too small, resulting in the network not learning at all. Later, the learning rate was adjusted to 1e-2, and both the train loss and validate loss of the network achieved rapid decline (And the optimizer is Adam). When adjusting the learning rate later, you can start from 1 to the minor, do not preconceive.
I'm playing with CIFAR-10 dataset using ResNet-50 on Keras with Tensorflow backend, but I ran into a very strange training pattern, where the model loss decreased first, and then started to increase until it plateaued/stuck at a single value due to almost 0 learning rate. Correspondingly, the model accuracy increased first, and then started to decrease until it plateaued at 10% (aka random guess). I wonder what is going wrong?
Typically this U shaped pattern happens with a learning rate that is too large (like this post), but it is not the case here. This pattern also doesn't' look like a classic "over-fitting" as both the training and validation loss increase over time. In the answer to the above linked post, someone mentioned that if Adam optimizer is used, loss may explode under small learning rate when local minimum is exceeded, I'm not sure I can follow what is said there, and also I'm using SGD with weight decay instead of Adam.
Specifically for the training set up, I used resent50 with random initialization, SGD optimizer with 0.9 momentum and a weight decay of 0.0001 using decoupled weight decay regularization, batch-size 64, initial learning_rate 0.01 which declined by a factor of 0.5 with each 10 epochs of non-decreasing validation loss.
base_model = tf.keras.applications.ResNet50(include_top=False,
weights=None,pooling='avg',
input_shape=(32,32,3))
prediction_layer = tf.keras.layers.Dense(10)
model = tf.keras.Sequential([base_model,
prediction_layer])
SGDW = tfa.optimizers.extend_with_decoupled_weight_decay(tf.keras.optimizers.SGD)
optimizer = SGDW(weight_decay=0.0001, learning_rate=0.01, momentum=0.9)
loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),
metrics=["accuracy"])
reduce_lr= tf.keras.callbacks.ReduceLROnPlateau(monitor='val_loss',factor=0.5, patience=10)
model.compile(optimizer=optimizer,
loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),
metrics=["accuracy"])
model.fit(train_batches, epochs=250,
validation_data=validation_batches,
callbacks=[reduce_lr])
I trained a network on a real-value labels (floating point numbers from 0.0 to 1.0) - several residual blocks in the beginning, and the last layers are
fully-connected layer with 64 neurons + ELU activation,
fully-connected layer with 16 neurons + ELU activation,
output logistic regression layer ( 1 neuron with y = 1 / (1 + exp(-x) ).
After training, I visualised weights of the layer with 16 neurons:
figure rows represents weights that every single 1 of 16 neurons developed for every single 1 of 64 neurons of previous layer, indices are 0..15 and 0..63;
UPD: figure shows neurons weights correlation (Pearson);
UPD: figure shows neurons weights MAD (mean absolute difference) - this proves redundancy event better than correlation.
Now the detailed questions:
Can we say that there are redundant features? I see several redundant groups of neurons: 0,4; 1,6,7 (maybe 8,11,15 too); 2,14; 12,13 (maybe) .
is it bad ?
if so, is there any regularizer, that penalizes redundant neuron weights, and makes neurons develop uncorrelated weights?
I use adam regularizer, Xavier initialization (the best of the tested), weight decay 1e-5/batch (the best of the tested), other output layers did not work as well as logistic regression (by means of precison & recall & lack of overfitting).
I use only 10 filters in each resnet blocks (which are 10, too) to address overfitting.
Are you using Tensorflow ? if yes, is post training quantization an option ?
tensorflow.org/lite/performance/post_training_quantization
This has some similar effect to what you need but also makes other improvements.
Alternatively maybe you can also try to use Quantization-aware training
https://github.com/tensorflow/tensorflow/tree/r1.14/tensorflow/contrib/quantize
While training a convolutional neural network following this article, the accuracy of the training set increases too much while the accuracy on the test set settles.
Below is an example with 6400 training examples, randomly chosen at each epoch (so some examples might be seen at the previous epochs, some might be new), and 6400 same test examples.
For a bigger data set (64000 or 100000 training examples), the increase in training accuracy is even more abrupt, going to 98 on the third epoch.
I also tried using the same 6400 training examples each epoch, just randomly shuffled. As expected, the result is worse.
epoch 3 loss 0.54871 acc 79.01
learning rate 0.1
nr_test_examples 6400
TEST epoch 3 loss 0.60812 acc 68.48
nr_training_examples 6400
tb 91
epoch 4 loss 0.51283 acc 83.52
learning rate 0.1
nr_test_examples 6400
TEST epoch 4 loss 0.60494 acc 68.68
nr_training_examples 6400
tb 91
epoch 5 loss 0.47531 acc 86.91
learning rate 0.05
nr_test_examples 6400
TEST epoch 5 loss 0.59846 acc 68.98
nr_training_examples 6400
tb 91
epoch 6 loss 0.42325 acc 92.17
learning rate 0.05
nr_test_examples 6400
TEST epoch 6 loss 0.60667 acc 68.10
nr_training_examples 6400
tb 91
epoch 7 loss 0.38460 acc 95.84
learning rate 0.05
nr_test_examples 6400
TEST epoch 7 loss 0.59695 acc 69.92
nr_training_examples 6400
tb 91
epoch 8 loss 0.35238 acc 97.58
learning rate 0.05
nr_test_examples 6400
TEST epoch 8 loss 0.60952 acc 68.21
This is my model (I'm using RELU activation after each convolution):
conv 5x5 (1, 64)
max-pooling 2x2
dropout
conv 3x3 (64, 128)
max-pooling 2x2
dropout
conv 3x3 (128, 256)
max-pooling 2x2
dropout
conv 3x3 (256, 128)
dropout
fully_connected(18*18*128, 128)
dropout
output(128, 128)
What could be the cause?
I'm using Momentum Optimizer with learning rate decay:
batch = tf.Variable(0, trainable=False)
train_size = 6400
learning_rate = tf.train.exponential_decay(
0.1, # Base learning rate.
batch * batch_size, # Current index into the dataset.
train_size*5, # Decay step.
0.5, # Decay rate.
staircase=True)
# Use simple momentum for the optimization.
optimizer = tf.train.MomentumOptimizer(learning_rate,
0.9).minimize(cost, global_step=batch)
This is very much expected. This problem is called over-fitting. This is when your model starts "memorizing" the training examples without actually learning anything useful for the Test set. In fact, this is exactly why we use a test set in the first place. Since if we have a complex enough model we can always fit the data perfectly, even if not meaningfully. The test set is what tells us what the model has actually learned.
Its also useful to use a Validation set which is like a test set, but you use it to find out when to stop training. When the Validation error stops lowering you stop training. why not use the test set for this? The test set is to know how well your model would do in the real world. If you start using information from the test set to choose things about your training process, than its like your cheating and you will be punished by your test error no longer representing your real world error.
Lastly, convolutional neural networks are notorious for their ability to over-fit. It has been shown the Conv-nets can get zero training error even if you shuffle the labels and even random pixels. That means that there doesn't have to be a real pattern for the Conv-net to learn to represent it. This means that you have to regularize a conv-net. That is, you have to use things like Dropout, batch normalization, early stopping.
I'll leave a few links if you want to read more:
Over-fitting, validation, early stopping
https://elitedatascience.com/overfitting-in-machine-learning
Conv-nets fitting random labels:
https://arxiv.org/pdf/1611.03530.pdf
(this paper is a bit advanced, but its interresting to skim through)
P.S. to actually improve your test accuracy you will need to change your model or train with data augmentation. You might want to try transfer learning as well.
I have two classes with 3 images each. I tried this code in Keras.
trainingDataGenerator = ImageDataGenerator()
trainGenerator = trainingDataGenerator.flow_from_directory(
trainingDataDir,
target_size=(28, 28),
batch_size = 1,
seed=7,
class_mode='binary',
)
FilterSize = (3,3)
inputShape = (imageWidth, imageHeight,3)
model = Sequential()
model.add (Conv2D(32, FilterSize, input_shape= inputShape))
model.add (Activation('relu'))
model.add ( MaxPooling2D(pool_size=(2,2)))
model.add(Flatten())
model.add(Activation('relu'))
model.add(Dense(1))
model.add(Activation('sigmoid'))
model.compile(loss='binary_crossentropy',
optimizer = 'rmsprop',
metrics=['accuracy'])
model.fit_generator(
trainGenerator,
steps_per_epoch=3,
epochs=epochs)
My Output:
When I train this model, I get this output:
Using TensorFlow backend.
Found 2 images belonging to 2 classes.
Epoch 1/1
3/3 [==============================] - 0s - loss: 5.3142 - acc: 0.6667
My Question:
I wonder how it determines the loss and accuracy and on what basis? (ie: loss: 5.3142 - acc: 0.6667 ). I have not given any validation image to validate the model to find accuracy and loss. Does this loss, and accuracy is against the input image itself?
In short, can we say something like this: "This model has accuracy of %, and loss of % without validation images"?
The training loss and accuracy is calculated not by comparing to validation data but rather by comparing the prediction of your neural network of sample x with the label y for that sample that you provide in your training set.
You initialize your neural network and (usually) set all weights to a random value with a certain deviation. After that you feed the features of your training dataset into your network, and let it "guess" the outcome aka the label that you have (if you do supervised learning like in your case).
Then your framework compares that guess with the actual label and calculates the error which it then backpropagates through your network thereby adjusting and improving all weights.
This works perfectly well without any validation data.
Validation data serves you to see the quality of your model (loss, accuracy etc.) by letting the model predict on unseen data. With that you get the so called validation loss / accuracy and with this information you tune your hyperparameters.
In a last step you use your test data to evaluate the final quality of your training.