"""Calculate accuracy.
Authors
* Jianyuan Zhong 2020
"""
import torch
from speechbrain.dataio.dataio import length_to_mask
def Accuracy(log_probabilities, targets, length=None):
"""Calculates the accuracy for predicted log probabilities and targets in a batch.
Arguments
---------
log_probabilities : torch.Tensor
Predicted log probabilities (batch_size, time, feature).
targets : torch.Tensor
Target (batch_size, time).
length : torch.Tensor
Length of target (batch_size,).
Returns
-------
numerator : float
The number of correct samples
denominator : float
The total number of samples
Example
-------
>>> probs = torch.tensor([[0.9, 0.1], [0.1, 0.9], [0.8, 0.2]]).unsqueeze(0)
>>> acc = Accuracy(torch.log(probs), torch.tensor([1, 1, 0]).unsqueeze(0), torch.tensor([2/3]))
>>> print(acc)
(1.0, 2.0)
"""
if length is not None:
mask = length_to_mask(
length * targets.shape[1],
max_len=targets.shape[1],
).bool()
if len(targets.shape) == 3:
mask = mask.unsqueeze(2).repeat(1, 1, targets.shape[2])
padded_pred = log_probabilities.argmax(-1)
if length is not None:
numerator = torch.sum(
padded_pred.masked_select(mask) == targets.masked_select(mask)
)
denominator = torch.sum(mask)
else:
numerator = torch.sum(padded_pred == targets)
denominator = targets.shape[1]
return float(numerator), float(denominator)
class AccuracyStats:
"""Module for calculate the overall one-step-forward prediction accuracy.
Example
-------
>>> probs = torch.tensor([[0.9, 0.1], [0.1, 0.9], [0.8, 0.2]]).unsqueeze(0)
>>> stats = AccuracyStats()
>>> stats.append(torch.log(probs), torch.tensor([1, 1, 0]).unsqueeze(0), torch.tensor([2/3]))
>>> acc = stats.summarize()
>>> print(acc)
0.5
"""
def __init__(self):
self.correct = 0
self.total = 0
def append(self, log_probabilities, targets, length=None):
"""This function is for updating the stats according to the prediction
and target in the current batch.
Arguments
---------
log_probabilities : torch.Tensor
Predicted log probabilities (batch_size, time, feature).
targets : torch.Tensor
Target (batch_size, time).
length : torch.Tensor
Length of target (batch_size,).
"""
numerator, denominator = Accuracy(log_probabilities, targets, length)
self.correct += numerator
self.total += denominator
def summarize(self):
"""Computes the accuracy metric."""
return self.correct / self.total