Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

saving & loading NeuralNet model

See original GitHub issue

hi matt & henning,

so i managed to use tagger.py from the examples directory with some easy patch (using train_sparse and predict_sparse directly), but i don’t know how to save and load the trained model. i’ve tried to re-initialised the tagger.model and set the weights but the final result is too random (i’m guessing re-init causing the model set to default/random).

In [2]: tagger, w, c = pickle.load(open('models/UD_Indonesian/model.bin'))

In [3]: tagger.model = NeuralNet([tagger.ex.input_length] + [tagger.hidden_width] * tagger.depth + [len(tagger.classes)], embed=(tagger.ex.tables, tagger.ex.slots), rho=tagger.L2, eta=tagger.learn_rate, update_step=tagger.solver)

In [4]: tagger.model.weights = w

thanks!

Issue Analytics

State:
Created 8 years ago
Comments:7 (4 by maintainers)

Top GitHub Comments

1reaction

honnibalcommented, Dec 31, 2016

Neural network is finally live 😃.

The new code is much simpler and better, and supports pickle — so serialisation is now no problem.

0reactions

honnibalcommented, Feb 24, 2016

I think youre clipping the value in that __set__ method — weight_t is double in recent versions, so be careful to use the typedef.

Top Results From Across the Web

How to Save and Load Your Keras Deep Learning Model

In this post, you will discover how to save your Keras models to files and load them up again to make predictions. After...

Save and load models | TensorFlow Core

Model progress can be saved during and after training. This means a model can resume where it left off and avoid long training...

Training, saving and loading Artificial Neural Networks in Keras

We demonstrate how to code a Artificial neural network model and train and save it in JSON or H5 format which can be...

Saving and Loading Models - Deeplearning4j

Saving and loading of neural networks. ... Utility class suited to save/restore neural net models. writeModel. public static void writeModel(@NonNull Model ...

Save and Load a Model with TensorFlow's Keras API

... the various ways of saving and loading a Sequential model using TensorFlow's Keras API. ... TensorFlow - Python Deep Learning Neural Network...