1
$\begingroup$

I have started to study ANNs with Tensorflow and Keras. Now I want to find a solution to use ANNs over Hadoop. I have learnt that Spark 2.0 does have a Multilayer Perceptron Classifier, but as far as I can see it is quite "primitive" in comparison with TF and Keras (there are only feedforward types and hidden layer/output layer activation functions are hard-wired), there is no wide variety of optimizers, cost functions, architecture types etc.

Is there any competitive alternative for large scale neural networks? Of course, I could use Amazon with very powerful, GPU-driven machines but these are not parallel frameworks either...

$\endgroup$
2
1
$\begingroup$

You might take a look over TensorFrames, a Databricks library which allow running TensorFlow code on top of Apache Spark :

https://github.com/databricks/tensorframes

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.