Metrics

You can record metrics from within tasks and workflows. Metrics are streamed in real-time and rendered as charts in Studio.

Recording metrics

Create a Metric and call record():

import coflux as cf

loss = cf.Metric("loss", group="training")
accuracy = cf.Metric("accuracy", group="training")

@cf.task()
def train(epochs):
    for epoch in range(epochs):
        loss_val, acc_val = run_epoch(epoch)
        loss.record(loss_val, at=epoch)
        accuracy.record(acc_val, at=epoch)

By default, values are associated with the time since the execution started. The at= parameter lets you specify an explicit x-value — useful when plotting against epochs, steps, or any other dimension.

Groups and scales

Metrics within the same group are displayed together on a shared chart. You can configure the axes with MetricGroup and MetricScale:

group = cf.MetricGroup("training", units="epoch")

loss = cf.Metric("loss", group=group, scale=cf.MetricScale(units=""))
accuracy = cf.Metric("accuracy", group=group, scale=cf.MetricScale(units="%", lower=0, upper=100))

MetricGroup configures the x-axis (shared across metrics in the group):

Parameter	Description
name	Group name (required)
units	X-axis unit label
lower	X-axis lower bound
upper	X-axis upper bound

MetricScale configures the y-axis. Metrics sharing a scale name within a group share a y-axis:

Parameter	Description
name	Scale name (metrics with the same name share a y-axis)
units	Y-axis unit label
lower	Y-axis lower bound
upper	Y-axis upper bound
progress	Whether to render as a progress bar

Progress

The progress() helper wraps an iterable and automatically records a progress metric as items are consumed:

@cf.task()
def process_batch(items):
    for item in cf.progress(items):
        handle(item)

This renders as a progress bar in Studio. You can customise the metric key and group:

for item in cf.progress(items, key="items_processed", group="processing"):
    handle(item)

Throttling

Metrics are throttled client-side to avoid flooding the server in tight loops. By default, each metric reports up to 10 values per second. This can be adjusted per metric:

metric = cf.Metric("fast_metric", throttle=100)   # 100 values/sec
metric = cf.Metric("exact_metric", throttle=None)  # No throttling

Metric store

Currently a built-in metric store is supported. The intention is to support integrating with external stores in future.