It supports feed-forward networks such as convolutional neural networks (CNNs), recurrent neural networks including Long short-term memory (LSTM) and any combination thereof . It allows architectures of multiple inputs and multiple outputs, including auxiliary classifiers . It ca be optimized using Nesterov momentum, RMSprop and ADAM . It is definable and does not need to derive gradients due to Theano's symbolic differentiation . It provides transparent support of CPUs and GPUs due to Theano's expression compiler.
It is utilized in defining, evaluating and optimizing mathematical expressions. The software provides easy to use layers to achieve high-level application programming interface (API). Lasagne can do everything Theano can and with added benefits of user-friendly layering functions.
Sander Dieleman, Jan Schlüter, Colin Raffel, Eben Olson, Søren Kaae Sønderby, Daniel Nouri and others
Documentaries, videos and podcasts
- Theano (software)A numerical computation library for Python.
- KerasKeras is an open source neural networks API written in Python built on top of Tensorflow, CNTK or Theano.
- TensorFlowA machine learning software library for numerical calculation
- Microsoft Cognitive Toolkit
- PyTorchA deep learning framework and an open source machine learning library for Python that provides Tensor computation with strong GPU acceleration and deep neural networks built on Autograd system
- Show More