CI (License MIT 1.0)


Use Keras models in C++ with ease

Table of contents


Would you like to build/train a model using Keras/Python? And would you like to run the prediction (forward pass) on your model in C++ without linking your application against TensorFlow? Then frugally-deep is exactly for you.


  • is a small header-only library written in modern and pure C++.
  • is very easy to integrate and use.
  • depends only on FunctionalPlus, Eigen and json – also header-only libraries.
  • supports inference (model.predict) not only for sequential models but also for computational graphs with a more complex topology, created with the functional API.
  • re-implements a (small) subset of TensorFlow, i.e., the operations needed to support prediction.
  • results in a much smaller binary size than linking against TensorFlow.
  • works out-of-the-box also when compiled into a 32-bit executable. (Of course, 64 bit is fine too.)
  • utterly ignores even the most powerful GPU in your system and uses only one CPU core per prediction. 😉
  • but is quite fast on one CPU core compared to TensorFlow, and you can run multiple predictions in parallel, thus utilizing as many CPUs as you like to improve the overall prediction throughput of your application/pipeline.

Supported layer types

Layer types typically used in image recognition/generation are supported, making many popular model architectures possible (see Performance section).

  • Add, Concatenate, Subtract, Multiply, Average, Maximum
  • AveragePooling1D/2D, GlobalAveragePooling1D/2D
  • Bidirectional, TimeDistributed, GRU, LSTM, CuDNNGRU, CuDNNLSTM
  • Conv1D/2D, SeparableConv2D, DepthwiseConv2D
  • Cropping1D/2D, ZeroPadding1D/2D
  • BatchNormalization, Dense, Flatten, Normalization
  • Dropout, AlphaDropout, GaussianDropout, GaussianNoise
  • SpatialDropout1D, SpatialDropout2D, SpatialDropout3D
  • RandomContrast, RandomFlip, RandomHeight
  • RandomRotation, RandomTranslation, RandomWidth, RandomZoom
  • MaxPooling1D/2D, GlobalMaxPooling1D/2D
  • ELU, LeakyReLU, ReLU, SeLU, PReLU
  • Sigmoid, Softmax, Softplus, Tanh
  • Exponential, GELU, Softsign
  • UpSampling1D/2D
  • Reshape, Permute, RepeatVector
  • Embedding

Also supported

  • multiple inputs and outputs
  • nested models
  • residual connections
  • shared layers
  • variable input shapes
  • arbitrary complex model architectures / computational graphs
  • custom layers (by passing custom factory functions to load_model)

Currently not supported are the following:

ActivityRegularization, AveragePooling3D, Conv2DTranspose (why), Conv3D, ConvLSTM2D, Cropping3D, Dot, GRUCell, LocallyConnected1D, LocallyConnected2D, LSTMCell, Masking, MaxPooling3D, RepeatVector, RNN, SimpleRNN, SimpleRNNCell, StackedRNNCells, ThresholdedReLU, Upsampling3D, temporal models


  1. Use Keras/Python to build (model.compile(...)), train (model.fit(...)) and test (model.evaluate(...)) your model as usual. Then save it to a single HDF5 file using model.save('....h5', include_optimizer=False). The image_data_format in your model must be channels_last, which is the default when using the TensorFlow backend. Models created with a different image_data_format and other backends are not supported.

  2. Now convert it to the frugally-deep file format with keras_export/convert_model.py

  3. Finally load it in C++ (fdeep::load_model(...)) and use model.predict(...) to invoke a forward pass with your data.

The following minimal example shows the full workflow:

# create_model.py
import numpy as np
from tensorflow.keras.layers import Input, Dense
from tensorflow.keras.models import Model

inputs = Input(shape=(4,))
x = Dense(5, activation='relu')(inputs)
predictions = Dense(3, activation='softmax')(x)
model = Model(inputs=inputs, outputs=predictions)
model.compile(loss='categorical_crossentropy', optimizer='nadam')

    np.asarray([[1, 2, 3, 4], [2, 3, 4, 5]]),
    np.asarray([[1, 0, 0], [0, 0, 1]]), epochs=10)

model.save('keras_model.h5', include_optimizer=False)
python3 keras_export/convert_model.py keras_model.h5 fdeep_model.json

<div class="highlight highlight-source-c++ position-relative overflow-auto" data-snippet-clipboard-copy-content="// main.cpp
int main()
const auto model = fdeep::load_model("fdeep_model.json");
const auto result = model.predict(
std::vector{1, 2, 3, 4})});
std::cout << fdeep::show_tensors(result) <

// main.cpp
#include <fdeep/fdeep.hpp>
int main()
    const auto model = fdeep::load_model("fdeep_model.json");
    const auto result = model.predict(
        std::vector<float>{1, 2, 3, 4})});
    std::cout << fdeep::show_tensors(result) << std::endl;