* \ingroup dim * \struct Dim * \brief The Dim struct stores information about the dimensionality of expressions. * \details Batch dimension is trea

* \ingroup operations * \brief Expressions are the building block of a Dynet computation graph * \details [long description] */

* \ingroup params * \brief Initialize parameters with a constant value */

* \ingroup tensor * \brief Represents a tensor of any order * \details This provides a bridge between classic C++ types and Eigen tensors. * */

* \ingroup params * \brief Initialize parameters with samples from a uniform distribution * */

* \ingroup params * \brief Initialize with the methods described in [Glorot, 2010](http://www.jmlr.org/proceedings/papers/v9/glorot10a/glorot10a.pdf?

* \ingroup rnnbuilders * @brief VanillaLSTM allows the creation of a "standard" LSTM, ie with decoupled input and forget gates and no peephole connec

if DYNET exhausts its memory pool

* Wrapper around Dict that also maintains unigram counts * * Stops counting when frozen */

* \ingroup ffbuilders * \struct Layer * \brief Simple layer structure * \details Contains all parameters defining a layer * */

* \ingroup params * \brief Initializes from a `std::vector` of floats */

--- Le Cun's uniform distribution * Initialize parameters with samples from a Le Cun's uniform distribution * Reference: LeCun 98, Efficient Backprop

* \ingroup rnnbuilders * \brief Builds a tree-LSTM which is recursively defined by a * Bidirectional LSTM over the node and its children representat

* \ingroup params * \brief Object representing a trainable lookup parameter * */

* \ingroup params * \brief Object representing a trainable parameter * \details This objects acts as a high level component linking the actual param

* \ingroup params * \brief Initialize as the identity * \details This will raise an exception if used on non square matrices */

* \ingroup params * \brief Initializes according to [Saxe et al., 2014](https://arxiv.org/abs/1312.6120) * \details Initializes as a random orthogon

`RNNBuilder` trait