When I compile my torch model to torchscript I can make use of the function forward by just calling the torchscript model object model().
But when I want to use another function created on the model I cant call the function. I try to do model.functionName() expecting to call the function functionName but nothing happens.
Any idea on how could I call subfunctions from the model object that are not forward?
Thanks
I suppose you should add decorator #torch.jit.export above the method you want to call and convert torchscript model again. After that you can call method by module.run_method(name, args) docs.
Related
Since Model.train_on_batch() doesn't take a callback input, I tried using Callback.on_epoch_end() in order to write my loss to tensorboard
However, trying to run the on_epoch_end() method results in the titular error, 'TensorBoard' object has no attribute 'writer'. Other solutions to my original problem with writing to tensorboard included calling the Callback.writer attribute, and running these solutions gave the same error. Also, the tensorflow documentation for the TensorBoard class doesn't mention a writer attribute
I'm somewhat of a novice programmer, but it seems to me that the on_epoch_end() method is also at some point calling the writer attribute, but I'm confused as to why the function would use an attribute that doesn't exist
Here's the code I'm using to create the callback:
logdir = "./logs/"
tensorboard_callback = tf.keras.callbacks.TensorBoard(log_dir=logdir)
and this is the callback code that I try to run in my training loop:
logs = {
'encoder':encoder_loss[0],
'discriminator':d_loss,
'generator':g_loss,
}
tensorboard_callback.on_epoch_end(i, logs)
where encoder_loss, d_loss, and g_loss are my scalars, and i is the batch number
Is the error a result of some improper code on my part, or is tensorflow trying to reference something that doesn't exist?
Also, if anyone knows another way to write to tensorboard using Model.train_on_batch, that would also solve my problem
Since you are using a callback without the fit method, you also need to pass your model to the TensorBoard object:
logdir = "./logs/"
tensorboard_callback = tf.keras.callbacks.TensorBoard(log_dir=logdir)
tensorboard_callback.set_model(model=model)
What all inbuilt function call happen after creating a object from class? I know init() get called any other functions get called automatically ?
I don't think anything occurs after init() unless you define it to, before init(); however a call to the class' __new__() function will occur BEFORE the call to init()
To load a DLL and call function in VC++, we use LoadLibrary and GetProcAddress , which needs mangled name of the method. Does it instantiate a class object and then call the method?
If it does not then how can I call a method in a class but instantiating an object of the class and then call a method?
What is process to load .h file, load class, instantiate object and then call a method in VC++
There is actually a very good example on CodeProject that describes exactly how to do this.
EDIT with reference to your comment, if you read the above article, you'll see that GetProcAddress() does nothing more than return a function pointer. If you want to create an instance of a class that's exported from the DLL, you have to allocate memory and force the system to invoke the constructor of the class to instantiate it there. Once that's done, however, I'm sort of assuming that you can call functions defined in the class in the normal way by using the object you just created.
I am writing unit tests in java using mockito.
This is the statement that I am trying to test.
final Map<EntityKey, Element<Movie>> resultMap = Watcher.watch(movies);
movies is Set of movie names which is a key to identify a movie.
I mocked watcher class
final Watcher<Movie> watcher = mock(Watcher.class);
Mockito.when(watcher.watch(Matchers.any(Set.class))).thenReturn()
what to include in "thenReturn"??
In the thenReturn function you need to pass an object of the same type as the method you are mocking's return type.
When this method is then called on that object, it will return the object you passed to thenReturn instead of actually going into the function.
This is the core concept behind mocking.
Having said that. If you are trying to test the Watcher.watch method then you probably don't want to mock it anyway. You should only mock those classes you are NOT testing.
You would need to make a Map<EntityKey,Element<Movie>> that would be suitable for use in the rest of the test. I'm not quite sure what your test is actually trying to assert, but whatever it is, choose the Map accordingly. Your Map object is what you want to return from thenReturn.
How can I test the initBefore method of Groovy Domain-Classes with a unit test in Grails?
I created the dummy object but the beforeInsert-method is not called until myObject.save() is invoked and save is unavailable in the testing environments.
Edit: its a unit-test. there is no error, but the method beforeInsert is not called
beforeInsert is called during unit tests. I can verify this in my tests. A couple things to consider:
ensure you use a beforeInsert method, and not a closure. A closure will not work correctly.
it is called when the object is flushed, so perhaps you are having silent save errors beforehand. In your test when you save the object do .save(flush: true, failOnError: true)
Do you want test if the beforeInsert method is being called or the logic of beforeInsert is correct?
If you want to test if beforeInsert is being called the test class should extend the GrailsUnitTestCase. Do so should give you mocking capabilities and add all the methods like save() and validate(). You can verify if the mocked object called the beforeInsert method or not when you do a save().
If you are testing the logic of beforeInsert then you don't need to mock it. You can create the object and test the logic just like other unit test.
Hope this helps.
Just creat a domain object and save() it. Then check whether or not the "beforeInsert" manipulated your Object.
save() is available in the testing enviroments. Please show your Stacktrace when calling myDomainobject.save()
I had the same exact problem! In GORM (at least until the current version) the save method does not take effect immediately just because it is called! If you want it to take effect right away you need to specify flush:true like this domain.save(flush:true).
it says here http://grails.org/doc/2.2.x/ref/Domain%20Classes/save.html
The save method informs the persistence context that an instance
should be saved or updated. The object will not be persisted
immediately unless the flush argument is used:
To answer your question, beforeInsert is not called until the save is persisted (save takes effect) therefor you should invoke save with flush to test beforeInsert and beforeUpdate methods.
Hope this helps!