# Keras Mixture Density Network Layer
[](https://github.com/cpmpercussion/keras-mdn-layer/actions/workflows/python-app.yml)
[](https://zenodo.org/badge/latestdoi/137585470)
[](https://badge.fury.io/py/keras-mdn-layer)
A mixture density network (MDN) Layer for Keras using TensorFlow's distributions module. This makes it a bit more simple to experiment with neural networks that predict multiple real-valued variables that can take on multiple equally likely values.
This layer can help build MDN-RNNs similar to those used in [RoboJam](https://github.com/cpmpercussion/robojam), [Sketch-RNN](https://experiments.withgoogle.com/sketch-rnn-demo), [handwriting generation](https://distill.pub/2016/handwriting/), and maybe even [world models](https://worldmodels.github.io). You can do a lot of cool stuff with MDNs!
One benefit of this implementation is that you can predict any number of real-values. TensorFlow's `Mixture`, `Categorical`, and `MultivariateNormalDiag` distribution functions are used to generate the loss function (the probability density function of a mixture of multivariate normal distributions with a diagonal covariance matrix). In previous work, the loss function has often been specified by hand which is fine for 1D or 2D prediction, but becomes a bit more annoying after that.
Two important functions are provided for training and prediction:
- `get_mixture_loss_func(output_dim, num_mixtures)`: This function generates a loss function with the correct output dimensiona and number of mixtures.
- `sample_from_output(params, output_dim, num_mixtures, temp=1.0)`: This functions samples from the mixture distribution output by the model.
## Installation
This project requires Python 3.6+, TensorFlow and TensorFlow Probability. You can easily install this package from [PyPI](https://pypi.org/project/keras-mdn-layer/) via `pip` like so:
python3 -m pip install keras-mdn-layer
And finally, import the module in Python: `import keras_mdn_layer as mdn`
Alternatively, you can clone or download this repository and then install via `python setup.py install`, or copy the `mdn` folder into your own project.
## Build
This project builds using `poetry`. To build a wheel use `poetry build`.
## Examples
Some examples are provided in the notebooks directory.
To run these using `poetry`, run `poetry install` and then open jupyter `poetry run jupyter lab`.
There's scripts for fitting multivalued functions, a standard MDN toy problem:
<img src="https://preview.ibb.co/mZzkpd/Keras_MDN_Demo.jpg" alt="Keras MDN Demo" border="0">
There's also a script for generating fake kanji characters:
<img src="https://i.ibb.co/yFvtgkL/kanji-mdn-examples.png" alt="kanji test 1" border="0" width="600"/>
And finally, for learning how to generate musical touch-screen performances with a temporal component:
<img src="https://i.ibb.co/WpzSCV8/robojam-examples.png" alt="Robojam Model Examples" border="0">
## How to use
The MDN layer should be the last in your network and you should use `get_mixture_loss_func` to generate a loss function. Here's an example of a simple network with one Dense layer followed by the MDN.
from tensorflow import keras
import keras_mdn_layer as mdn
N_HIDDEN = 15 # number of hidden units in the Dense layer
N_MIXES = 10 # number of mixture components
OUTPUT_DIMS = 2 # number of real-values predicted by each mixture component
model = keras.Sequential()
model.add(keras.layers.Dense(N_HIDDEN, batch_input_shape=(None, 1), activation='relu'))
model.add(mdn.MDN(OUTPUT_DIMS, N_MIXES))
model.compile(loss=mdn.get_mixture_loss_func(OUTPUT_DIMS,N_MIXES), optimizer=keras.optimizers.Adam())
model.summary()
Fit as normal:
history = model.fit(x=x_train, y=y_train)
The predictions from the network are parameters of the mixture models, so you have to apply the `sample_from_output` function to generate samples.
y_test = model.predict(x_test)
y_samples = np.apply_along_axis(sample_from_output, 1, y_test, OUTPUT_DIMS, N_MIXES, temp=1.0)
See the notebooks directory for examples in jupyter notebooks!
### Load/Save Model
Saving models is straight forward:
model.save('test_save.h5')
But loading requires `cutom_objects` to be filled with the MDN layer, and a loss function with the appropriate parameters:
m_2 = keras.models.load_model('test_save.h5', custom_objects={'MDN': mdn.MDN, 'mdn_loss_func': mdn.get_mixture_loss_func(1, N_MIXES)})
## Acknowledgements
- Hat tip to [Omimo's Keras MDN layer](https://github.com/omimo/Keras-MDN) for a starting point for this code.
- Super hat tip to [hardmaru's MDN explanation, projects, and good ideas for sampling functions](http://blog.otoro.net/2015/11/24/mixture-density-networks-with-tensorflow/) etc.
- Many good ideas from [Axel Brando's Master's Thesis](https://github.com/axelbrando/Mixture-Density-Networks-for-distribution-and-uncertainty-estimation)
- Mixture Density Networks in Edward [tutorial](http://edwardlib.org/tutorials/mixture-density-network).
## References
1. Christopher M. Bishop. 1994. Mixture Density Networks. [Technical Report NCRG/94/004](http://publications.aston.ac.uk/373/). Neural Computing Research Group, Aston University. http://publications.aston.ac.uk/373/
2. Axel Brando. 2017. Mixture Density Networks (MDN) for distribution and uncertainty estimation. Master’s thesis. Universitat Politècnica de Catalunya.
3. A. Graves. 2013. Generating Sequences With Recurrent Neural Networks. ArXiv e-prints (Aug. 2013). https://arxiv.org/abs/1308.0850
4. David Ha and Douglas Eck. 2017. A Neural Representation of Sketch Drawings. ArXiv e-prints (April 2017). https://arxiv.org/abs/1704.03477
5. Charles P. Martin and Jim Torresen. 2018. RoboJam: A Musical Mixture Density Network for Collaborative Touchscreen Interaction. In Evolutionary and Biologically Inspired Music, Sound, Art and Design: EvoMUSART ’18, A. Liapis et al. (Ed.). Lecture Notes in Computer Science, Vol. 10783. Springer International Publishing. DOI:[10.1007/9778-3-319-77583-8_11](http://dx.doi.org/10.1007/9778-3-319-77583-8_11)
Raw data
{
"_id": null,
"home_page": "https://github.com/cpmpercussion/keras-mdn-layer",
"name": "keras-mdn-layer",
"maintainer": null,
"docs_url": null,
"requires_python": "==3.11.*",
"maintainer_email": null,
"keywords": "mixture density layer, neural network, machine learning",
"author": "Charles Martin",
"author_email": "cpm@charlesmartin.au",
"download_url": "https://files.pythonhosted.org/packages/26/cf/a95a0b51f21f3c04f8cd17a1adb9204c835610255ae59781fc1bddb21c90/keras_mdn_layer-0.4.4.tar.gz",
"platform": null,
"description": "# Keras Mixture Density Network Layer\n\n[](https://github.com/cpmpercussion/keras-mdn-layer/actions/workflows/python-app.yml)\n\n[](https://zenodo.org/badge/latestdoi/137585470)\n[](https://badge.fury.io/py/keras-mdn-layer)\n\nA mixture density network (MDN) Layer for Keras using TensorFlow's distributions module. This makes it a bit more simple to experiment with neural networks that predict multiple real-valued variables that can take on multiple equally likely values.\n\nThis layer can help build MDN-RNNs similar to those used in [RoboJam](https://github.com/cpmpercussion/robojam), [Sketch-RNN](https://experiments.withgoogle.com/sketch-rnn-demo), [handwriting generation](https://distill.pub/2016/handwriting/), and maybe even [world models](https://worldmodels.github.io).\u00a0You can do a lot of cool stuff with MDNs!\n\nOne benefit of this implementation is that you can predict any number of real-values. TensorFlow's `Mixture`, `Categorical`, and `MultivariateNormalDiag` distribution functions are used to generate the loss function (the probability density function of a mixture of multivariate normal distributions with a diagonal covariance matrix). In previous work, the loss function has often been specified by hand which is fine for 1D or 2D prediction, but becomes a bit more annoying after that.\n\nTwo important functions are provided for training and prediction:\n\n- `get_mixture_loss_func(output_dim, num_mixtures)`: This function generates a loss function with the correct output dimensiona and number of mixtures.\n- `sample_from_output(params, output_dim, num_mixtures, temp=1.0)`: This functions samples from the mixture distribution output by the model.\n\n## Installation \n\nThis project requires Python 3.6+, TensorFlow and TensorFlow Probability. You can easily install this package from [PyPI](https://pypi.org/project/keras-mdn-layer/) via `pip` like so:\n\n python3 -m pip install keras-mdn-layer\n\nAnd finally, import the module in Python: `import keras_mdn_layer as mdn`\n\nAlternatively, you can clone or download this repository and then install via `python setup.py install`, or copy the `mdn` folder into your own project.\n\n## Build\n\nThis project builds using `poetry`. To build a wheel use `poetry build`.\n\n## Examples\n\nSome examples are provided in the notebooks directory.\n\nTo run these using `poetry`, run `poetry install` and then open jupyter `poetry run jupyter lab`.\n\nThere's scripts for fitting multivalued functions, a standard MDN toy problem:\n\n<img src=\"https://preview.ibb.co/mZzkpd/Keras_MDN_Demo.jpg\" alt=\"Keras MDN Demo\" border=\"0\">\n\nThere's also a script for generating fake kanji characters:\n\n<img src=\"https://i.ibb.co/yFvtgkL/kanji-mdn-examples.png\" alt=\"kanji test 1\" border=\"0\" width=\"600\"/>\n\nAnd finally, for learning how to generate musical touch-screen performances with a temporal component:\n\n<img src=\"https://i.ibb.co/WpzSCV8/robojam-examples.png\" alt=\"Robojam Model Examples\" border=\"0\">\n\n## How to use\n\nThe MDN layer should be the last in your network and you should use `get_mixture_loss_func` to generate a loss function. Here's an example of a simple network with one Dense layer followed by the MDN.\n\n from tensorflow import keras\n import keras_mdn_layer as mdn\n\n N_HIDDEN = 15 # number of hidden units in the Dense layer\n N_MIXES = 10 # number of mixture components\n OUTPUT_DIMS = 2 # number of real-values predicted by each mixture component\n\n model = keras.Sequential()\n model.add(keras.layers.Dense(N_HIDDEN, batch_input_shape=(None, 1), activation='relu'))\n model.add(mdn.MDN(OUTPUT_DIMS, N_MIXES))\n model.compile(loss=mdn.get_mixture_loss_func(OUTPUT_DIMS,N_MIXES), optimizer=keras.optimizers.Adam())\n model.summary()\n\nFit as normal:\n\n history = model.fit(x=x_train, y=y_train)\n\nThe predictions from the network are parameters of the mixture models, so you have to apply the `sample_from_output` function to generate samples.\n\n y_test = model.predict(x_test)\n y_samples = np.apply_along_axis(sample_from_output, 1, y_test, OUTPUT_DIMS, N_MIXES, temp=1.0)\n\nSee the notebooks directory for examples in jupyter notebooks!\n\n### Load/Save Model\n\nSaving models is straight forward:\n\n model.save('test_save.h5')\n\nBut loading requires `cutom_objects` to be filled with the MDN layer, and a loss function with the appropriate parameters:\n\n m_2 = keras.models.load_model('test_save.h5', custom_objects={'MDN': mdn.MDN, 'mdn_loss_func': mdn.get_mixture_loss_func(1, N_MIXES)})\n\n\n## Acknowledgements\n\n- Hat tip to [Omimo's Keras MDN layer](https://github.com/omimo/Keras-MDN) for a starting point for this code.\n- Super hat tip to [hardmaru's MDN explanation, projects, and good ideas for sampling functions](http://blog.otoro.net/2015/11/24/mixture-density-networks-with-tensorflow/) etc.\n- Many good ideas from [Axel Brando's Master's Thesis](https://github.com/axelbrando/Mixture-Density-Networks-for-distribution-and-uncertainty-estimation)\n- Mixture Density Networks in Edward [tutorial](http://edwardlib.org/tutorials/mixture-density-network).\n\n## References\n\n1. Christopher M. Bishop. 1994. Mixture Density Networks. [Technical Report NCRG/94/004](http://publications.aston.ac.uk/373/). Neural Computing Research Group, Aston University. http://publications.aston.ac.uk/373/\n2. Axel Brando. 2017. Mixture Density Networks (MDN) for distribution and uncertainty estimation. Master\u2019s thesis. Universitat Polit\u00e8cnica de Catalunya.\n3. A. Graves. 2013. Generating Sequences With Recurrent Neural Networks. ArXiv e-prints (Aug. 2013). https://arxiv.org/abs/1308.0850\n4. David Ha and Douglas Eck. 2017. A Neural Representation of Sketch Drawings. ArXiv e-prints (April 2017). https://arxiv.org/abs/1704.03477\n5. Charles P. Martin and Jim Torresen. 2018. RoboJam: A Musical Mixture Density Network for Collaborative Touchscreen Interaction. In Evolutionary and Biologically Inspired Music, Sound, Art and Design: EvoMUSART \u201918, A. Liapis et al. (Ed.). Lecture Notes in Computer Science, Vol. 10783. Springer International Publishing. DOI:[10.1007/9778-3-319-77583-8_11](http://dx.doi.org/10.1007/9778-3-319-77583-8_11)\n",
"bugtrack_url": null,
"license": "MIT",
"summary": "An MDN Layer for Keras using TensorFlow's distributions module",
"version": "0.4.4",
"project_urls": {
"Homepage": "https://github.com/cpmpercussion/keras-mdn-layer",
"Repository": "https://github.com/cpmpercussion/keras-mdn-layer"
},
"split_keywords": [
"mixture density layer",
" neural network",
" machine learning"
],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "fb0d2d469d7d5735e79d0c53af2e6d7a7fb2f1584b2aade5eb2cf0cd1e13bf4a",
"md5": "cf6630c0d1f8afae418e9c4d24c8a783",
"sha256": "0820873089640481611301d2b19f2fe2953f792d3acd23ceb35c525470a650f9"
},
"downloads": -1,
"filename": "keras_mdn_layer-0.4.4-py3-none-any.whl",
"has_sig": false,
"md5_digest": "cf6630c0d1f8afae418e9c4d24c8a783",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": "==3.11.*",
"size": 9160,
"upload_time": "2024-07-23T06:26:19",
"upload_time_iso_8601": "2024-07-23T06:26:19.920337Z",
"url": "https://files.pythonhosted.org/packages/fb/0d/2d469d7d5735e79d0c53af2e6d7a7fb2f1584b2aade5eb2cf0cd1e13bf4a/keras_mdn_layer-0.4.4-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "26cfa95a0b51f21f3c04f8cd17a1adb9204c835610255ae59781fc1bddb21c90",
"md5": "bd6ddbab5db8810a383f77c912137076",
"sha256": "d668311f9e9f524533f564a86d4f2f97c0d72bd5eefc259a56cf0937c79e6453"
},
"downloads": -1,
"filename": "keras_mdn_layer-0.4.4.tar.gz",
"has_sig": false,
"md5_digest": "bd6ddbab5db8810a383f77c912137076",
"packagetype": "sdist",
"python_version": "source",
"requires_python": "==3.11.*",
"size": 8095,
"upload_time": "2024-07-23T06:26:21",
"upload_time_iso_8601": "2024-07-23T06:26:21.330809Z",
"url": "https://files.pythonhosted.org/packages/26/cf/a95a0b51f21f3c04f8cd17a1adb9204c835610255ae59781fc1bddb21c90/keras_mdn_layer-0.4.4.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2024-07-23 06:26:21",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "cpmpercussion",
"github_project": "keras-mdn-layer",
"travis_ci": false,
"coveralls": false,
"github_actions": true,
"lcname": "keras-mdn-layer"
}
JSON
