noggin.logger.LiveMetric

class noggin.logger.LiveMetric(name: str)

Holds the relevant data for a train/test metric for live plotting.

Attributes:

batch_data

Batch-level measurements of the metric.

batch_domain

Array of iteration-counts at which the metric was recorded.

epoch_data

Epoch-level measurements of the metrics.

epoch_domain

Array of iteration-counts at which an epoch was set for this metric.

name

Name of the metric.

Methods

add_datapoint(value, weighting)	Record a batch-level measurement of the metric.
from_dict(metrics_dict, numpy.ndarray])	The inverse of LiveMetric.to_dict.
set_epoch_datapoint(x)	Mark the present iteration as an epoch, and compute the mean value of the metric since the past epoch.
to_dict()	Returns the batch data, epoch domain, and epoch data in a dictionary.

__init__(name: str)

Parameters:

name : str

Raises:

TypeError

Invalid metric name (must be string)

Methods

__init__(name)	Parameters:
add_datapoint(value, weighting)	Record a batch-level measurement of the metric.
from_dict(metrics_dict, numpy.ndarray])	The inverse of LiveMetric.to_dict.
set_epoch_datapoint(x)	Mark the present iteration as an epoch, and compute the mean value of the metric since the past epoch.
to_dict()	Returns the batch data, epoch domain, and epoch data in a dictionary.

Attributes

batch_data	Batch-level measurements of the metric.
batch_domain	Array of iteration-counts at which the metric was recorded.
epoch_data	Epoch-level measurements of the metrics.
epoch_domain	Array of iteration-counts at which an epoch was set for this metric.
name	Name of the metric.