nnet-descriptor.h File Reference

This file contains class definitions for classes ForwardingDescriptor, SumDescriptor and Descriptor. More...

#include "base/kaldi-common.h"
#include "util/kaldi-io.h"
#include "matrix/matrix-lib.h"
#include "nnet3/nnet-common.h"
#include "nnet3/nnet-component-itf.h"
#include <iostream>
#include <sstream>
#include <vector>
#include <map>
Include dependency graph for nnet-descriptor.h:
This graph shows which files directly or indirectly include this file:

Go to the source code of this file.


class  ForwardingDescriptor
 A ForwardingDescriptor describes how we copy data from another NetworkNode, or from multiple other NetworkNodes, possibly with a scalar weight. More...
class  SimpleForwardingDescriptor
 SimpleForwardingDescriptor is the base-case of ForwardingDescriptor, consisting of a source node in the graph with a given scalar weight (which will in the normal case be 1.0). More...
class  OffsetForwardingDescriptor
 Offsets in 't' and 'x' values of other ForwardingDescriptors. More...
class  SwitchingForwardingDescriptor
 Chooses from different inputs based on the the time index modulo (the number of ForwardingDescriptors given as inputs). More...
class  RoundingForwardingDescriptor
 For use in clockwork RNNs and the like, this forwarding-descriptor rounds the time-index t down to the the closest t' <= t that is an exact multiple of t_modulus_. More...
class  ReplaceIndexForwardingDescriptor
 This ForwardingDescriptor modifies the indexes (n, t, x) by replacing one of them (normally t) with a constant value and keeping the rest. More...
class  SumDescriptor
 This is an abstract base-class. More...
class  OptionalSumDescriptor
 This is the case of class SumDescriptor, in which we contain just one term, and that term is optional (an IfDefined() expression). More...
class  SimpleSumDescriptor
 This is the normal base-case of SumDescriptor which just wraps a ForwardingDescriptor. More...
class  ConstantSumDescriptor
 This is an alternative base-case of SumDescriptor (an alternative to SimpleSumDescriptor) which represents a constant term, e.g. More...
class  BinarySumDescriptor
 BinarySumDescriptor can represent either A + B, or (A if defined, else B). More...
class  Descriptor
struct  GeneralDescriptor
 This class is only used when parsing Descriptors. More...


 This code computes Goodness of Pronunciation (GOP) and extracts phone-level pronunciation feature for mispronunciations detection tasks, the reference:

Detailed Description

This file contains class definitions for classes ForwardingDescriptor, SumDescriptor and Descriptor.

Basically this is code that specifies how we glue together the outputs of possibly several other network-nodes, as the input of a particular network node (or as an output of the network). In the neural-network code we refer to the top-level descriptor which is Descriptor. The InputDescriptor is a concatenation of features; each part is a SumDescriptor. The SumDescriptor is a summation over a set of features of all the same dimension, each of which is represented by a ForwardingDescriptor. A ForwardingDescriptor in the simplest case just takes just points you to a particular network node, but in general can do things like adding time offsets, and selecting different rows of its matrix from different inputs. Unlike the other descriptors, a ForwardingDescriptor is in general a bit like a parse tree, in that it can in general contain other ForwardingDescriptors.

The following gives an overview of the expressions that can appear in descriptors. Caution; this is a simplification that overgenerates descriptors: not all combinations are allowed.

<descriptor>  ::=   <node-name>      ;; node name of kInput or kComponent node.
<descriptor>  ::=   Append(<descriptor>, <descriptor> [, <descriptor> ... ] )
<descriptor>  ::=   Sum(<descriptor>, <descriptor>)
<descriptor>  ::=   Const(<value>, <dimension>)    ;; e.g. Const(1.0, 512)
<descriptor>  ::=   Scale(<scale>, <descriptor>)   ;; e.g. Scale(-1.0, tdnn2)
;; Failover or IfDefined might be useful for time t=-1 in a RNN, for instance.
<descriptor>  ::=   Failover(<descriptor>, <descriptor>)   ;; 1st arg if computable, else 2nd
<descriptor>  ::=   IfDefined(<descriptor>)     ;; the arg if defined, else zero.
<descriptor>  ::=   Offset(<descriptor>, <t-offset> [, <x-offset> ] ) ;; offsets are integers
;; Switch(...) is intended to be used in clockwork RNNs or similar schemes.  It chooses
;; one argument based on the value of t (in the requested Index) modulo the number of
;; arguments
<descriptor>  ::=   Switch(<descriptor>, <descriptor> [, <descriptor> ...])
;; For use in clockwork RNNs or similar, Round() rounds the time-index t of the
;; requested Index to the next-lowest multiple of the integer <t-modulus>,
;; and evaluates the input argument for the resulting Index.
<descriptor>  ::=   Round(<descriptor>, <t-modulus>)  ;; <t-modulus> is an integer
;; ReplaceIndex replaces some <variable-name> (t or x) in the requested Index
;; with a fixed integer <value>.  E.g. might be useful when incorporating
;; iVectors; iVector would always have time-index t=0.
<descriptor>  ::=   ReplaceIndex(<descriptor>, <variable-name>, <value>)

Definition in file nnet-descriptor.h.