Copies parameters and buffers from :attr:`state_dict` into this module and its descendants. If :attr:`strict` is ``True`` then the key

Returns the set of category names.

Which split contains item i.

Returns the number of distinct labels (plus zero), i.e., one more than the maximum label number. If a category is specified,

Iterator for all batches

Composes two stacked fieldmap maps. Field maps are represented as triples of (offset, size, step), where each is an (x, y) pair. To find

Input: set of rows with 'number' fields (or another field name key). Output: array such that a[number] = the row with the given number.

3x3 convolution with padding

Returns the pixel coverage of the label in units of whole-images.

Extracts all elements from the TopN. This is a destructive operation. The only method that can be called immediately after extract() is reset

Returns a single batch as an array of dictionaries.

Given a 0, 1, 2, or 3-dimensional label_data and a default shape of the form (1, y, x), returns a 3d tensor by

Pushes a new element.

Computes the receptive field of a specific location. Parameters ---------- location: tuple The x-y position of the unit being que

Like numpy.zoom, but does not crash when the first dimension of the array is of size 1, as happens often with segmentations

Verify that we are using the correct data

Turn an answer into a vector

Turn a question into a vector of indices and a question length

Computes the precision@k for the specified values of k

Computes the precision@k for the specified values of k

Apply any number of attention maps over the input. The attention map has to have the same size in all dimensions except dim=1.

Runs beam search caption generation on a single image. Args: initial_state: An initial state for the recurrent model Retur

Input: set of rows with 'number' fields (or another field name key). Output: array such that a[number] = the row with the given number.

Iterator for batches as arrays of tensors.

The filename of the ith jpeg (original image).