pytorch-symbolic 0.7.4

Provides symbolic API for model creation in PyTorch.

Pytorch Symbolic

Pytorch Symbolic is MIT licensed library that adds symbolic API for model creation to PyTorch.

Defining complex models in PyTorch requires creating classes and writing boilerplate code.

Defining models in Keras is easier. Pytorch Symbolic makes it just as easy.

Features:

• Small extension of PyTorch
• No dependencies besides PyTorch
• Produces models entirely compatible with PyTorch
• Overhead free, tested in benchmarks
• Reduces the amount of code you write
• Works well with complex architectures
• Code and documentation is automatically tested

Example

To create a symbolic model, you'll use symbolic tensors and nn.Modules. Add layers by calling layer(symbolic_tensor) or equivalently symbolic_tensor(layer). That's all!

Create symbolic model just like in Keras: treat symbolic tensors as if they were normal tensors.

Layers will be automagically registered in your model.

from torch import nn
from pytorch_symbolic import Input

x = Input(shape=(1, 28, 28))  # Input is a SymbolicTensor
print(x)

x = nn.Flatten()(x)  # Every layer operation returns another SymbolicTensor
x = x(nn.Flatten())  # This is equivalent to previous line
print(x)

<SymbolicTensor at 0x7faf5779d130; child of 0; parent of 0>
<SymbolicTensor at 0x7fafea899f10; child of 1; parent of 0>

Using symbolic tensors, we can define a working classificator in a few lines of code:

from torch import nn
from pytorch_symbolic import Input, SymbolicModel

inputs = Input(shape=(1, 28, 28))
x = nn.Flatten()(inputs)
x = nn.Linear(x.shape[1], 10)(x)(nn.ReLU())
model = SymbolicModel(inputs=inputs, outputs=x)
model

SymbolicModel(
  (module0_depth1): Flatten(start_dim=1, end_dim=-1)
  (module1_depth2): Linear(in_features=784, out_features=10, bias=True)
  (module2_depth3): ReLU()
)

See more examples in Quick Start.

Gentle introduction

There's a notebook showing the basic usage of Pytorch Symbolic. You will:

• Learn Pytorch Symbolic in an interactive way
• Try the package before installing it on your computer
• See visualizations of graphs that are createdunder the hood

Click:

Installation

Install easily with pip:

pip install pytorch-symbolic

Links

• See Documentation
• See on GitHub
• See on PyPI

Create an issue if you noticed a problem! Send me an e-mail if you want to get involved: [e-mail].