Functional deep learning library with NumPy and PyTorch integration
A refreshing functional take on deep learning, compatible with your favorite libraries
$ pip install thinc$ python -c "from thinc.api import chain, with_array; model = chain(Dense(10), Relu(), Dense(5))"$ python -c "from thinc.api import Model; model.initialize(X=train_data); predictions = model.predict(test_data)"$ python -c "from thinc.layers import PyTorchWrapper; wrapped = PyTorchWrapper(pytorch_layer)"$ python -c "import thinc; print(thinc.__version__)"