torcheval.metrics.Throughput

class torcheval.metrics.Throughput(*, device: Optional[device] = None)

Calculate the throughput value which is the number of elements processed per second.

Note: In a distributed setting, it’s recommended to use world_size * metric.compute() to get an approximation of total throughput. While using sync_and_compute(metric) requires state sync. Additionally, sync_and_compute(metric) will give a slightly different value compared to world_size * metric.compute().

Examples:

>>> import time
>>> import torch
>>> from torcheval.metrics import Throughput
>>> metric = Throughput()
>>> items_processed = 64
>>> ts = time.monotonic()
>>> time.sleep(2.0)  # simulate executing the program for 2 seconds
>>> elapsed_time_sec = time.monotonic() - ts
>>> metric.update(items_processed, elapsed_time_sec)
>>> metric.compute()
tensor(32.)

__init__(*, device: Optional[device] = None) → None

Initialize a metric object and its internal states.

Use self._add_state() to initialize state variables of your metric class. The state variables should be either torch.Tensor, a list of torch.Tensor, or a dictionary with torch.Tensor as values

Methods

__init__(*[, device])	Initialize a metric object and its internal states.
compute()	Implement this method to compute and return the final metric value from state variables.
load_state_dict(state_dict[, strict])	Loads metric state variables from state_dict.
merge_state(metrics)	Implement this method to update the current metric's state variables to be the merged states of the current metric and input metrics.
reset()	Reset the metric state variables to their default value.
state_dict()	Save metric state variables in state_dict.
to(device, *args, **kwargs)	Move tensors in metric state variables to device.
update(num_processed, elapsed_time_sec)	Update states with the values and weights.

Attributes

device

The last input device of Metric.to().