Expression conv1d_narrow(const Expression& x, const Expression& f) { return Expression(x.pg, x.pg->add_function<Conv1DNarrow>({x.i, f.i})); } Expressi

Returns the final output of each hidden layer.

Returns the final state of each hidden layer.

returns the binary representation (in an int) of a float

Builds a `DynetParams` from command line arguments.

Returns an Expression representing a vector the size of the number of classes.

Returns the logits (before application of the softmax).

Get function name, ie removes possible return type

Returns the RNN state that is the parent of the given state.

* \brief Get the unique graph ID * \details This ID is incremented by 1 each time a computation graph is * created \return graph id */

Check if there is a parent class

* \brief Return the negative log likelihood for the (batched) pair (x,y) * \details For a batched input \f$\{x_i\}_{i=1,\dots,N}\f$, \f$\{y_i\}_{i=

---

* \brief Get the weight decay lambda value. */

Returns the L2 norm of the gradient in the ParameterCollection.

Returns gradients of the parameter as a tensor.

Inputs vector/matrix/tensor on the specified device.

Inputs scalar.

Inputs scalar on the specified device.

Inputs sparse vector on the specified device. # Arguments g - Computation graph. d - Dimension of the input matrix. ids - The indexes of the data po

Clears caches.

---

---

Load a lookup parameter from a model.

---

Load a parameter from a model.

Looks up parameter.

Make a grid of images, via numpy. Args: tensor (Tensor or list): 4D mini-batch Tensor of shape (B x C x H x W) or a list of i

Make a network graph, represented as of nodes and a set of edges. The nodes are represented as tuples: (name: string, input_dim: Dim, label: st

Download and extract processed data and embeddings.

These operations are implemented using FFTs (described below), but the slow quadratic implementation is left here for a reference.

Returns the number of components in `h_0`.

Creates batched one hot vectors on the specified device. # Arguments g - Computation graph. d - Dimension of the input vector. ids - The indices we

* \ingroup arithmeticoperations * \brief Matrix-scalar multiplication * \details Multiply an expression component-wise by a scalar. * * \param x T

* \ingroup arithmeticoperations * \brief Matrix-scalar division * \details Divide an expression component-wise by a scalar. * * \param x The matri

Returns the total number of tunable parameters.

Picks batch element

Picks batch elements

Load a lookup parameter.