iterative/vscode-dvc

# Monitor your Experiments

As you run your experiments you can use the same
[`Plots Dashboard`](command:dvc.showPlots) to watch for the metrics and plot
updates:

<p align="center">
  <img src="images/live-plots-updates.gif"
       alt="Live Metric Updates" />
</p>

This functionality comes baked in when you update your plots files in your
training code. For examples like this, use the
[DVCLive](https://dvc.org/doc/dvclive) library that also
[supports](https://dvc.org/doc/dvclive/ml-frameworks) many popular ML frameworks
to automate this process:

```python
from dvclive import Live

live = Live("training_metrics")

for epoch in range(NUM_EPOCHS):
    train_model(...)
    metrics = evaluate_model(...)

    for metric_name, value in metrics.items():
        live.log(metric_name, value)

    live.next_step()
```

`DVCLive` is _optional_, and you can just append or modify plot files using any
language and any tool.

💡 Plots created in the `Custom` section of the plots dashboard are being
updated automatically based on the data in the table. You don't even have to
manage or write any special plot files.