Running the examples in their own project (maven) about djl HOT 9 CLOSED

Self-contained example in https://github.com/salamanders/superdeep

from djl.

@salamanders
osx doesn't have GPU support the classifier should be: osx-x86_64

from djl.

Whohoo, that did it, thank you!

For any future readers, my pom.xml had:

        <dependency>
            <groupId>ai.djl</groupId>
            <artifactId>api</artifactId>
            <version>0.2.1</version>
        </dependency>
        <dependency>
            <groupId>ai.djl.mxnet</groupId>
            <artifactId>mxnet-model-zoo</artifactId>
            <version>0.2.1</version>
        </dependency>
        <dependency>
            <groupId>ai.djl.mxnet</groupId>
            <artifactId>mxnet-native-mkl</artifactId>
            <version>1.6.0-b</version>
            <classifier>osx-x86_64</classifier>
        </dependency>

from djl.

... so should I (ever) include mxnet-native-cu101mkl or mxnet-native-cu92mkl?

from djl.

You might. Each of the mxnet-native-* dependencies represent a different way of building the MXNet engine. For OSX, we only distribute the build with Intel MKL Library enabled (-mkl). For linux, we have more builds including builds with GPU support (this is your -cu101mkl and -cu92mkl). You can look at https://github.com/awslabs/djl/blob/master/mxnet/mxnet-engine/README.md for a list of the various builds that we distribute for each platform.

from djl.

I think there may be a typo, it says "macOS
For macOS, you can choose between the following two libraries:" and then it lists one.

from djl.

@salamanders
macOS is official os name since 10.12, OSX is legacy name.

from djl.

Sure! I'm happy with either.
I meant more that it says "following two options" but there is only one option.

from djl.

@salamanders
Thanks for clarify, I will submit a fix to address this.

from djl.