I was doing some research on training deep neural networks using tensorflow. I know how to train a model. My problem is i have to train the same model on 2 different computers with different datasets. Then save the model weights. Later i have to merge the 2 model weight files somehow. I have no idea how to merge them. Is there a function that does this or should the weights be averaged?
Any help on this problem would be useful
Thanks in advance
There is literally no way to merge weights, you cannot average or combine them in any way, as the result will not mean anything. What you could do instead is combine predictions, but for that the training classes have to be the same.
This is not a programming limitation but a theoretical one.
It is better to merge weight updates (gradients) during the training and keep a common set of weights rather than trying to merge the weights after individual trainings have completed. Both individually trained networks may find a different optimum and e.g. averaging the weights may give a network which performs worse on both datasets.
There are two things you can do:
Look at 'data parallel training': distributing forward and backward passes of the training process over multiple compute nodes each of which has a subset of the entire data.
In this case typically:
each node propagates a minibatch forward through the network
each node propagates the loss gradient backwards through the network
a 'master node' collects gradients from minibatches on all nodes and updates the weights correspondingly
and distributes the weight updates back to the compute nodes to make sure each of them has the same set of weights
(there are variants of the above to avoid that compute nodes idle too long waiting for results from others). The above assumes that Tensorflow processes running on the compute nodes can communicate with each other during the training.
Look at https://www.tensorflow.org/deploy/distributed) for more details and an example of how to train networks over multiple nodes.
If you really have train the networks separately, look at ensembling, see e.g. this page: https://mlwave.com/kaggle-ensembling-guide/ . In a nutshell, you would train individual networks on their own machines and then e.g. use an average or maximum over the outputs of both networks as a combined classifier / predictor.
Related
I have five classes and I want to compare four of them against one and the same class. This isn't a One vs Rest classifier, as for each output I want to score them against one base class.
The four outputs should be: base class vs classA, base class vs classB, etc.
I could do this by having multiple binary classification tasks, but that's wasting computation time if the first layers are BERT preprocessing + pretrained BERT layers, and the only differences between the four classifiers are the last few layers of BERT (finetuned ones) and the Dense layer.
So why not merge the graphs for more performance?
My inputs are four different datasets, each annotated with true/false for each class.
As I understand it, I can re-use most of the pipeline (BERT preprocessing and the first layers of BERT), as those have shared weights. I should then be able to train the last few layers of BERT and the Dense layer on top differently depending on the branch of the classifier (maybe using something like keras.switch?).
I have tried many alternative options including multi-class and multi-label classifiers, with actual and generated (eg, machine-annotated) labels in the case of multiple input labels, different activation and loss functions, but none of the results were acceptable to me (none were as good as the four separate models).
Is there a solution for merging the four different models for more performance, or am I stuck with using 4x binary classifiers?
When you train DNN for specific task it will be (in vast majority of cases) be better than the more general model that can handle several task simultaneously. Saying that, based on my experience the properly trained general model produces very similar results to the original binary ones. Anyways, here couple of suggestions for training strategies (assuming your training datasets for each task are completely different):
Weak supervision approach
Train your binary classifiers, and label your datasets using them (i.e. label with binary classifier trained on dataset 2 datasets [1,3,4]). Then train your joint model as multilabel task using all the newly labeled datasets (don't forget to randomize samples before feeding them to trainer ;) ). Here you will need to experiment if you will use threshold and set a label to 0/1 or use the scores of the binary classifiers.
Create custom loss function that will not penalize if no information provided for certain class. So when your will introduce sample from (say) dataset 2, your loss will be calculated only for the 2nd class.
Of course you can apply both simultaneously. For example, if you know that binary classifier produces scores that are polarized (most results are near 0 or 1), you can use weak labels, and automatically label your data with scores. Now during the second stage penalize loss such that for score x' = 4(x-0.5)^2 (note that you get logits from the model, so you will need to apply sigmoid function). This way you will increase contribution of the samples binary classifier is confident about, and reduce that of less certain ones.
As for releasing last layers of BERT, usually unfreezing upper 3-6 layers is enough. Releasing more layers improves results very little and increases time and memory requirements.
I have a feed forward network with batch normalization layers in it. the network has to be run multiple times before one backpropogation can be applied. It is used to compute:
a. vanilla logits (vl)
b. logits with noisy inputs (nl)
difference between vl and nl is also part of the cost function
I cant understand how and when to run tf.GraphKeys.UPDATE_OPS. If i run it typically by grouping it with optimizer i think the moving averages will be updated with mean and standard deviation of noisy batches since that is the last run of the network but i need the updates based on vanilla (no noise) run of the network. I can't change the order of different runs of the network as noise is computed using the output of vanilla run of the network
Usually, when using Keras, the datasets used to train the neural network are labeled.
For example, if I have a 100,000 rows of patients with 12 field per each row, then the last field will indicate if this patient is diabetic or no (0 or 1).
And then after training is finished I can insert a new record and predict if this person is diabetic or no.
But in the case of unlabeled datasets, where I can not label the data due to some reasons, how can I train the neural network to let him know that those are the normal records and any new record that does not match this network will be malicious or not accepted ?
This is called one-class learning and is usually done by using autoencoders. You train an autoencoder on the training data to reconstruct the data itself. The labels in this case is the input itself. This will give you a reconstruction error. https://en.wikipedia.org/wiki/Autoencoder
Now you can define a threshold where the data is benign or not, depending on the reconstruction error. The hope is that the reconstruction of the good data is better than the reconstruction of the bad data.
Edit to answer the question about the difference in performance between supervised and unsupervised learning.
This cannot be said with any certainty, because I have not tried it and I do not know what the final accuracy is going to be. But for a rough estimate supervised learning will perform better on the trained data, because more information is supplied to the algorithm. However if the actual data is quite different to the training data the network will underperform in practice, while the autoencoder tends to deal better with different data. Additionally, per rule of thumb you should have 5000 examples per class to train a neural network reliably, so labeling could take some time. But you will need some data to test anyways.
It sounds like you need fit two different models:
a model for bad record detection
a model for prediction of a patient's likelihood to be diabetic
For both of these models, you will need to have labels. For the first model your labels would indicate whether the record is good or bad (malicious) and the second would be whether the patient is diabetic or not.
In order to detect bad records, you may find that simple logistic regression or SVM performs adequately.
When I first train an LSTM in Keras on sequence data - my training data -
and then use model.predict() to make predictions with my test data as input, is the hidden state of the LSTM still being adjusted?
Basic operation of a neural network is to take an input (vector) which is connected to the output with connections and, sometimes, other layers such as context layers. These connections are modelled as matrices and vary in strength, we call these weight matrices.
This means that the only thing we do when we are feeding data into the network is to put a vector into the network, multiply the values with the weight matrix and call that the output. In special cases, like recurrent networks, we even keep some values stored in other vectors and combine this stored value with the current input.
During training we not only feed data into the network, we also compute an error value that we evaluate in a clever way so that it tells us how we should change the weight matrices we multiply our inputs (and possibly past inputs for recurrent layers) with.
Therefore: yes, of course the basic execution behavior does not change for recurrent layers. We are just not updating weights anymore.
There are layers that do behave differently during execution time because they are treated as regularisers, i.e. methods that make training the network more efficient, which are deemed as unnecessary during execution. Examples for these layers are Noise and BatchNormalization. Almost all neural network layers (including recurrent ones) include drop-out which is another form of regularisation which disables a random percentage of connections in the layer. This is also only done during training.
I'm using tensorflow to run a cnn for image classification.
I use tensorflow cifar10 cnn implementation.(tensorflow cifar10)
I want to decrease the number of connections, meaning I want to prune the low-weight connections.
How can I create a new graph(subgraph) without some of the nuerones?
Tensorflow does not allow you lock/freeze a particular kernel of a particular layer, that I have found. The only I've found to do this is to use the tf.assign() function as shown in
How to freeze/lock weights of one Tensorflow variable (e.g., one CNN kernel of one layer
It's fairly cave-man but I've seen no other solution that works. Essentially, you have to .assign() the values every so often as you iterate through the data. Since this approach is so inelegant and brute-force, it's very slow. I do the .assign() every 100 batches.
Someone please post a better solution and soon!
The cifar10 model you point to, and for that matter, most models written in TensorFlow, do not model the weights (and hence, connections) of individual neurons directly in the computation graph. For instance, for fully connected layers, all the connections between the two layers, say, with M neurons in the layer below, and 'N' neurons in the layer above, are modeled by one MxN weight matrix. If you wanted to completely remove a neuron and all of its outgoing connections from the layer below, you can simply slice out a (M-1)xN matrix by removing the relevant row, and multiply it with the corresponding M-1 activations of the neurons.
Another way is add an addition mask to control the connections.
The first step involves adding mask and threshold variables to the
layers that need to undergo pruning. The variable mask is the same
shape as the layer's weight tensor and determines which of the weights
participate in the forward execution of the graph.
There is a pruning implementation under tensorflow/contrib/model_pruning to prune the model. Hope this can help you to prune model quickly.
https://github.com/tensorflow/tensorflow/tree/master/tensorflow/contrib/model_pruning
I think google has an updated answer here : https://github.com/tensorflow/tensorflow/tree/master/tensorflow/contrib/model_pruning
Removing pruning nodes from the trained graph:
$ bazel build -c opt contrib/model_pruning:strip_pruning_vars
$ bazel-bin/contrib/model_pruning/strip_pruning_vars --checkpoint_path=/tmp/cifar10_train --output_node_names=softmax_linear/softmax_linear_2 --filename=cifar_pruned.pb
I suppose that cifar_pruned.pb will be smaller, since the pruned "or zero masked" variables are removed.