Saving uboost BDT with tf/keras base estimators

[srishtibhasin]

Hi,

I am trying to use a uBoost BDT to achieve uniform signal efficiency. My base estimator is a Keras model (Tensorflow 2.2), which I have written as a scikit-learn BaseEstimator subclass using tensorflow.keras.wrappers.scikit_learn.KerasClassifier. The training and everything seems to work fine, but I am encountering an error when I try to save the uboost classifier with pickle/joblib. The error is TypeError: can't pickle _thread.RLock objects (full error at bottom - it is mostly a long thread of calls to pickle )

From trying to look it up it seems the error is usually to do with the way tensorflow is run, but I'm only creating a simple model and fitting and all the session handling should be taken care of in this version of tf/keras. Maybe this answer is related https://github.com/keras-team/keras/issues/8343#issuecomment-470376703 ie. perhaps there is a call to something from the model that leaves an unserializable tensor object? As I am using the BDT not the classifier, I assume it is not to do with any parallel processes either?

Please let me know if you know what is causing the issue or if there is some way I can work around it.

Thanks!

[arogozhnikov]

Poor picklability of keras is a long-known issue (you can google keras with the same mistake error).

You may be fortunate to have some of variables being passed e.g. in lambda not through calls, https://stackoverflow.com/questions/55280201/keras-typeerror-cant-pickle-thread-rlock-objects

But otherwise I'm not sure there will be a simple solution

[srishtibhasin]

Yes, but of course keras has its own save_model functionality. So do you think it is simply not possible to save a uBoost model which is based on a keras model?

[arogozhnikov]

it is simply not possible to save a uBoost model which is based on a keras model?

It all goes down do pickle-ability of items. hep_ml's uBoost is completely pickle-able, but keras model is not.

Option 1. Ask keras maintainers why your model is not pickle-able Option 2. Store keras models separately from uBoost.

estimators = clf.estimators_ # list of keras models
# TODO save estimators somehow using keras tools
# delete estimators 
clf.estimators_ = None
with open('uboost.pkl', 'wb') as f:
   joblib.dump(clf, f)

# loading
with open('uboost.pkl', 'rb') as f:
   clf = joblib.load(f)

estimators = .... # load  estimators
clf.estimators_ = estimators

[arogozhnikov]

there is option 3 as well - find truly sklearn-compatible NN package =)

[srishtibhasin]

Ok I will try to implement option 2, thanks!