In multi-trial NAS, a sampled model should be able to be executed on a remote machine or a training platform (e.g., AzureML, OpenPAI). “Serialization” enables re-instantiation of model evaluator in another process or machine, such that, both the model and its model evaluator should be correctly serialized. To make NNI correctly serialize model evaluator, users should apply
nni.trace on some of their functions and objects. API references can be found in
Serialization is implemented as a combination of json-tricks and cloudpickle. Essentially, it is json-tricks, that is a enhanced version of Python JSON, enabling handling of serialization of numpy arrays, date/times, decimal, fraction and etc. The difference lies in the handling of class instances. Json-tricks deals with class instances with
__class__, which in most of our cases are not reliable (e.g., datasets, dataloaders). Rather, our serialization deals with class instances with two methods:
If the class / factory that creates the object is decorated with
nni.trace, we can serialize the class / factory function, along with the parameters, such that the instance can be re-instantiated.
Otherwise, cloudpickle is used to serialize the object into a binary.
The recommendation is, unless you are absolutely certain that there is no problem and extra burden to serialize the object into binary, always add
nni.trace. In most cases, it will be more clean and neat, and enables possibilities such as mutation of parameters (will be supported in future).
What will happen if I forget to “trace” my objects?
It is likely that the program can still run. NNI will try to serialize the untraced object into a binary. It might fail in complex cases. For example, when the object is too large. Even if it succeeds, the result might be a substantially large object. For example, if you forgot to add
MNIST, the MNIST dataset object wil be serialized into binary, which will be dozens of megabytes because the object has the whole 60k images stored inside. You might see warnings and even errors when running experiments. To avoid such issues, the easiest way is to always remember to add
nni.trace to non-primitive objects.
In NAS, serializer will throw exception when one of an single object in the recursive serialization is larger than 64 KB when binary serialized. This indicates that such object needs to be wrapped by
nni.trace. In rare cases, if you insist on pickling large data, the limit can be overridden by setting an environment variable
PICKLE_SIZE_LIMIT, whose unit is byte. Please note that even if the experiment might be able to run, this can still cause performance issues and even the crash of NNI experiment.
To trace a function or class, users can use decorator like,
@nni.trace class MyClass: ...
Inline trace that traces instantly on the object instantiation or function invoke is also acceptable:
Assuming a class
cls is already traced, when it is serialized, its class type along with initialization parameters will be dumped. As the parameters are possibly class instances (if not primitive types like
str), their serialization will be a similar problem. We recommend decorate them with
nni.trace as well. In other words,
nni.trace should be applied recursively if necessary.
Below is an example,
MNIST are serialized manually using
nni.trace takes a class / function as its argument, and returns a wrapped class and function that has the same behavior with the original class / function. The usage of the wrapped class / function is also identical to the original one, except that the arguments are recorded. No need to apply
pl.DataLoader because they are already traced.
import nni import nni.nas.evaluator.pytorch.lightning as pl from torchvision import transforms def create_mnist_dataset(root, transform): return MNIST(root='data/mnist', train=False, download=True, transform=transform) transform = nni.trace(transforms.Compose)([nni.trace(transforms.ToTensor)(), nni.trace(transforms.Normalize)((0.1307,), (0.3081,))]) # If you write like following, the whole transform will be serialized into a pickle. # This actually works fine, but we do NOT recommend such practice. # transform = transforms.Compose([transforms.ToTensor(), transforms.Normalize((0.1307,), (0.3081,))]) train_dataset = nni.trace(MNIST)(root='data/mnist', train=True, download=True, transform=transform) test_dataset = nni.trace(create_mnist_dataset)('data/mnist', transform=transform) # factory is also acceptable evaluator = pl.Classification(train_dataloaders=pl.DataLoader(train_dataset, batch_size=100), val_dataloaders=pl.DataLoader(test_dataset, batch_size=100), max_epochs=10)