A lightweight deep learning framework in Rust and Python
Cranberry is a deep learning framework developed using Rust and Python. It combines the performance of Rust with the flexibility of Python to create a powerful yet concise tool for machine learning practitioners and researchers.
- Minimal set of kernels implemented in Rust
- Majority of the framework logic implemented in Python for flexibility and ease of development
- Extremely concise codebase, making it easy to understand and extend
- PyTorch-like API for familiar and intuitive usage
- Designed for real-world applications while maintaining simplicity
Cranberry is in its early stages of development but is already functional for basic deep learning tasks. We are actively working on expanding its capabilities to make it a robust and practical framework for a wide range of applications.
git clone https://github.com/manoflearning/cranberry.git
cd cranberry
pip install poetry
poetry installfrom cranberry import nn, optim
from cranberry.features.datasets import mnist
X_train, Y_train, X_test, Y_test = mnist()
X_train, X_test = X_train.flatten(1), X_test.flatten(1)
model = nn.Sequential(nn.Linear(784, 128), nn.ReLU(), nn.Linear(128, 64), nn.ReLU(), nn.Linear(64, 10))
optimizer = optim.SGD(model.parameters(), lr=0.001)
for i in range(100):
optimizer.zero_grad()
# forward
loss = model(X_train).sparse_categorical_crossentropy(Y_train)
# backward
loss.backward()
# update
optimizer.step()
print(f"epoch {i}, loss {loss.item():.4f}")For more examples, see the examples directory.
We welcome contributions to help Cranberry grow into a fully-featured, production-ready deep learning framework. Also, join our Discord server to discuss development, get help, or just hang out with us.