Y = a*X+b.
.cudnnAlgoMode(ConvolutionLayer.AlgoMode.PREFER_FASTEST)configuration on convolution layers
RnnOutputLayer) labels. Otherwise, use 'c' ordered arrays. This is for faster memory access.
Xmxto? That depends on how much RAM is on your computer. In general, allocate as much heap space as you think the JVM will need to get work done. Let’s say you’re on a 16G RAM laptop — allocate 8G of RAM to the JVM. A sound minimum on laptops with less RAM would be 3g, so
Xmx. If they are unequal, the JVM will progressively allocate more memory as needed until it reaches the max, and that process of gradual allocation slows things down. You want to pre-allocate it at the beginning. So
.bash_profilefile, which adds environmental variables to bash. To see those variables, enter
envin the command line. To add more heap space, enter this command in your console:
DatasetIteratoryou create takes another 8 bytes.
Xmxand works extensively with off-heap memory. The off-heap memory will not surpass the amount of heap space you specify.
dl4j-examplesrepo, we don't make the ETL asynchronous, because the point of examples is to keep them simple. But for real-world problems, you need asynchronous ETL, and we'll show you how to do it with examples.
DatasetIteratorhides the complexity of loading data on disk. The code for using any Datasetiterator will always be the same, invoking looks the same, but they work differently.
AsyncDataSetIteratoris what you would use most of the time. It's described in the Javadoc here.
AsyncMultiDataSetIterator, described in the Javadoc here.
prefetchSizeis another parameter to set. Normal batch size might be 1000 examples, but if you set
prefetchSizeto 3, it would pre-fetch 3,000 instances.
fit, you're recreating a dataset, over and over again. We allow it to happen for ease of use, but we can show you how to speed things up. There are ways to make it just as fast.
Recordreaderdatasetiteratortalks to Datavec and outputs datasets for DL4J.