# Revision history [back]

known bug

known bug , unfortunately no way to do so without falling back to JNI

known bug , unfortunately no way to do so without falling back to JNIJNI.

you'll have to create a new EM instance with: : Ptr<ml::EM> em = Algorithm::load<ml::EM>(filename);

and return that to your java code

known bug , unfortunately no way to do so without falling back to JNI.

you'll have to create a new EM instance with: :

Ptr<ml::EM> em = Algorithm::load<ml::EM>(filename);Algorithm::load<ml::EM>(filename);


and return that to your java code

known bug , unfortunately no way to do so without falling back to JNI.

you'll have to create a new EM instance with:

Ptr<ml::EM> em = Algorithm::load<ml::EM>(filename);


and return that to your java code

(and again, the problem with the script, that generates the java/python/matlab wrappers is, that it cannot deduce the template argument properly)

known bug , unfortunately no way to do so without falling back to JNI.

you'll have to create a new EM instance (in c++) with:

Ptr<ml::EM> em = Algorithm::load<ml::EM>(filename);


and return that to your java code

(and again, the problem with the script, that generates the java/python/matlab wrappers is, that it cannot deduce the template argument properly)

known bug , unfortunately no way to do so without falling back to JNI.

you'll have to create a new EM instance (in c++) with:

Ptr<ml::EM> em = Algorithm::load<ml::EM>(filename);


and return that to your java codecode like:

    return (jlong)(new Ptr<ml::EM>(em));


(and again, the problem with the script, that generates the java/python/matlab wrappers is, that it cannot deduce the template argument properly)

that's a known bug , unfortunately no way to do so without falling back to JNI.

you'll have to create a new EM instance (in c++) with:

Ptr<ml::EM> em = Algorithm::load<ml::EM>(filename);


and return that to your java code like:

    return (jlong)(new Ptr<ml::EM>(em));


(and again, the problem with the script, that generates the java/python/matlab wrappers is, that it cannot deduce the template argument properly)

that's a known bug , unfortunately no way to do so without falling back to JNI.

you'll have to create a new EM instance (in c++) with:

Ptr<ml::EM> em = Algorithm::load<ml::EM>(filename);


and return that to your java code like:

    return (jlong)(new Ptr<ml::EM>(em));


yes, it has to be wrapped into a 2nd Ptr to keep the refcounts proper ;(

(and again, the problem with the script, that generates the java/python/matlab wrappers is, that it cannot deduce the template argument properly)

that's a known bug , for now, unfortunately there's no way to do so without falling back to JNI.

you'll have to create a new EM instance (in c++) with:

Ptr<ml::EM> em = Algorithm::load<ml::EM>(filename);


and return that to your java code like:

    return (jlong)(new Ptr<ml::EM>(em));


yes, it has to be wrapped into a 2nd Ptr to keep the refcounts proper ;(

(and again, the problem with the script, that generates the java/python/matlab wrappers is, that it cannot deduce the template argument properly)

that's a known bug , for now, unfortunately there's no way to do so without falling back to JNI.

you'll have to create a new EM instance (in c++) with:

Ptr<ml::EM> em = Algorithm::load<ml::EM>(filename);


and return that to your java code like:

    return (jlong)(new Ptr<ml::EM>(em));


yes, it has to be wrapped into a 2nd Ptr to keep the refcounts proper ;(

(and again, the problem with the script, that generates the java/python/matlab wrappers is, that it cannot deduce the template argument properly)argument)

that's a known bug , for now, unfortunately there's no way to do so without falling back to JNI.

you'll have to create a new EM instance (in c++) with:

Ptr<ml::EM> em = Algorithm::load<ml::EM>(filename);


and return that to your java code like:

    return (jlong)(new Ptr<ml::EM>(em));


yes, it has to be wrapped into a 2nd Ptr to keep the refcounts proper ;(

(and again, the problem with the script, that generates the java/python/matlab wrappers is, that it cannot deduce the template argument)argument at that stage)

that's a known bug , for now, unfortunately there's no way to do so without falling back to JNI.JNI/NDK.

you'll have to create a new EM instance (in c++) with:

Ptr<ml::EM> em = Algorithm::load<ml::EM>(filename);


and return that to your java code like:

    return (jlong)(new Ptr<ml::EM>(em));


yes, it has to be wrapped into a 2nd Ptr to keep the refcounts proper ;(

(and again, the problem with the script, that generates the java/python/matlab wrappers is, that it cannot deduce the template argument at that stage)

that's a known bug , for now, unfortunately there's no way to do so without falling back to JNI/NDK.JNI/NDK (the problem is not specific to EM, but applies to all the ml classes).

you'll have to create a new EM instance (in c++) with:

Ptr<ml::EM> em = Algorithm::load<ml::EM>(filename);


and return that to your java code like:

    return (jlong)(new Ptr<ml::EM>(em));


yes, it has to be wrapped into a 2nd Ptr to keep the refcounts proper ;(

(and again, the problem with the script, that generates the java/python/matlab wrappers is, that it cannot deduce the template argument at that stage)

that's a known bug , for now, unfortunately there's no way to do so without falling back to JNI/NDK (the problem is not even specific to EM, but applies to all the ml classes).

you'll have to create a new EM instance (in c++) with:

Ptr<ml::EM> em = Algorithm::load<ml::EM>(filename);


and return that to your java code like:

    return (jlong)(new Ptr<ml::EM>(em));


yes, it has to be wrapped into a 2nd Ptr to keep the refcounts proper ;(

(and again, the problem with the script, that generates the java/python/matlab wrappers is, that it cannot deduce the template argument at that stage)