pytorch2keras

Pytorch to Keras model convertor. Still beta for now.

Installation

pip install pytorch2keras 

Important notice

In that moment the only PyTorch 0.2 (deprecated) and PyTorch 0.4 (latest stable) are supported.

To use the converter properly, please, make changes in your ~/.keras/keras.json:

...
"backend": "tensorflow",
"image_data_format": "channels_first",
...

From the latest releases, multiple inputs is also supported.

Tensorflow.js

For the proper convertion to the tensorflow.js format, please use a new flag short_names=True.

How to build the latest PyTorch

Please, follow this guide to compile the latest version.

How to use

It's a convertor of pytorch graph to a Keras (Tensorflow backend) graph.

Firstly, we need to load (or create) pytorch model:

class TestConv2d(nn.Module):
    """Module for Conv2d convertion testing
    """

    def __init__(self, inp=10, out=16, kernel_size=3):
        super(TestConv2d, self).__init__()
        self.conv2d = nn.Conv2d(inp, out, stride=(inp % 3 + 1), kernel_size=kernel_size, bias=True)

    def forward(self, x):
        x = self.conv2d(x)
        return x

model = TestConv2d()

# load weights here
# model.load_state_dict(torch.load(path_to_weights.pth))

The next step - create a dummy variable with correct shapes:

input_np = np.random.uniform(0, 1, (1, 10, 32, 32))
input_var = Variable(torch.FloatTensor(input_np))

We're using dummy-variable in order to trace the model.

from converter import pytorch_to_keras
# we should specify shape of the input tensor
k_model = pytorch_to_keras(model, input_var, [(10, 32, 32,)], verbose=True)  

That's all! If all is ok, the Keras model is stores into the k_model variable.

Supported layers

Layers:

• Linear
• Conv2d
• Conv3d
• ConvTranspose2d
• MaxPool2d
• MaxPool3d
• AvgPool2d
• Global average pooling (as special case of AdaptiveAvgPool2d)
• Embedding
• UpsamplingNearest2d

Reshape:

• View
• Reshape (only with 0.4)
• Transpose (only with 0.4)

Activations:

• ReLU
• LeakyReLU
• PReLU (only with 0.2)
• SELU (only with 0.2)
• Tanh
• Softmax
• Softplus (only with 0.2)
• Softsign (only with 0.2)
• Sigmoid

Element-wise:

• Addition
• Multiplication
• Subtraction

Misc:

• reduce sum ( .sum() method)

Unsupported parameters

• Pooling: count_include_pad, dilation, ceil_mode
• Convolution: group

Models converted with pytorch2keras

• ResNet18
• ResNet34
• ResNet50
• SqueezeNet (with ceil_mode=False)
• DenseNet
• AlexNet
• Inception (v4 only)
• SeNet

Usage

Look at the tests directory.

License

This software is covered by MIT License.