https://github.com/sjmikler/pytorch-symbolic
Symbolic API for model creation in PyTorch.
https://github.com/sjmikler/pytorch-symbolic
neural-networks symbolic-execution
Last synced: 2 months ago
JSON representation
Symbolic API for model creation in PyTorch.
- Host: GitHub
- URL: https://github.com/sjmikler/pytorch-symbolic
- Owner: sjmikler
- License: mit
- Created: 2021-03-09T22:21:01.000Z (over 4 years ago)
- Default Branch: main
- Last Pushed: 2025-03-16T09:35:12.000Z (4 months ago)
- Last Synced: 2025-05-07T13:04:05.827Z (2 months ago)
- Topics: neural-networks, symbolic-execution
- Language: Jupyter Notebook
- Homepage:
- Size: 5.39 MB
- Stars: 66
- Watchers: 4
- Forks: 3
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE.txt
Awesome Lists containing this project
README
# Pytorch Symbolic
[//]: # (To get badges go to https://shields.io/ and use https://pypi.org/pypi/slicemap/json as data url. Query fields using dot as the separator.)
[](https://pypi.org/project/pytorch-symbolic)
[](https://github.com/sjmikler/pytorch-symbolic/blob/main/LICENSE.txt)
[](https://github.com/gahaalt/pytorch-symbolic/actions/workflows/notebook.yaml)
Pytorch Symbolic is MIT licensed library that adds symbolic API for model creation to PyTorch.
Pytorch Symbolic makes it easier and faster to define complex models.
It spares you writing boilerplate code.
It aims to be PyTorch equivalent for [Keras Functional API](https://keras.io/guides/functional_api/).
Features:
* Small extension of PyTorch
* No dependencies besides PyTorch
* Produces models entirely compatible with PyTorch
* Overhead free as tested in [benchmarks](docs/benchmarks.md)
* Reduces the amount of boilerplate code
* Works well with complex architectures
* Code and documentation is automatically tested
## Example
To create a symbolic model, you need Symbolic Tensors and `torch.nn.Module`.
Register layers and operations in your model by calling ``layer(inputs)`` or
equivalently ``inputs(layer)``.
Layers will be automagically added to your model and
all operations will be replayed on the real data.
That's all!
Using Pytorch Symbolic, we can define a working classifier in a few lines of code:
```python
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.Softmax(1))
model = SymbolicModel(inputs=inputs, outputs=x)
model.summary()
```
```stdout
_______________________________________________________
Layer Output shape Params Parent
=======================================================
1 Input_1 (None, 1, 28, 28) 0
2 Flatten_1 (None, 784) 0 1
3 Linear_1 (None, 10) 7850 2
4* Softmax_1 (None, 10) 0 3
=======================================================
Total params: 7850
Trainable params: 7850
Non-trainable params: 0
_______________________________________________________
```
**See more examples
in [Documentation Quick Start](docs/quick_start.md).**
## How to start
See Jupyter Notebook showing the basic usage of Pytorch Symbolic:
* Learn Pytorch Symbolic in an interactive way.
* Try the package before installing it on your computer.
* See visualizations of graphs that are created under the hood.
Open in Colab:
[](https://colab.research.google.com/github/gahaalt/pytorch-symbolic/blob/develop/introduction.ipynb)
## Installation
Install Pytorch Symbolic easily with pip:
```bash
pip install pytorch-symbolic
```
## Troubleshooting
Please create an issue if you notice a problem!
## Links
* [See Documentation](https://github.com/sjmikler/pytorch-symbolic/tree/main/docs)
* [See on GitHub](https://github.com/gahaalt/pytorch-symbolic/)
* [See on PyPI](https://pypi.org/project/pytorch-symbolic/)