Easy-to-use, one-liner AI model conversion tool between AI frameworks to ensure faster model development and portability with ONNX.
Project description
NeuralBridge
Easy-to-use, one-liner AI model conversion tool between AI frameworks to ensure faster model development and portability with Open Neural Network Exchange (ONNX).
How to Install
To install NeuralBridge simple use:
pip install neuralbridge
In case you would like to build this project from source please consult the requirements.txt
file and the src
directory.
How to Use
The main use case of NeuralBridge is to convert model runtime. It converts TensorFlow and PyTorch models from one framework to another absolutely automatically. You can access runtime conversion like this:
from neuralbridge import convert_model
model_in_other_framework = convert_model(model_in_one_framework)
Examples
You can find additional examples in the examples
directory.
Converting PyTorch model into TensorFlow
This example is in the example_tf_to_torch.py
file.
Converting TensorFlow model into PyTorch
This example is in the example_torch_to_tf.py
file.
Old design patterns
You can find the old design patterns in the following files:
-
old_way_onnx_to_tf.py
-
old_way_onnx_to_torch.py
-
old_way_tf_to_onnx.py
-
old_way_tf_to_torch.py
-
old_way_torch_to_onnx.py
-
old_way_torch_to_tf.py
Requirements
The list of all requirements are in the requirements.txt.
License
See details in the LICENSE.
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for neuralbridge-1.0.2-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | e361eb5b478bac88d560e5e2e4de4d1c4770da50dc61b311cd70b54f2289cead |
|
MD5 | ecab42df4f8330d124e958b5e7541edd |
|
BLAKE2b-256 | 2dd7e41bef2c5f06954c9e3447edac9a31b712f2d811dc2e6461136c3b241ece |