Explore ML Experiments

When you create a view by connecting to a Git repository, the view is added to your DVC Studio dashboard. You can dive deep into the views shown in your DVC Studio dashboard to explore all the ML experiments, visualize and compare them, and run new experiments. For this, open the view by clicking the view name (in this case, example-get-started).

A table will be generated as shown below. This includes metrics, hyperparameters and information about the datasets. All these values are flattened and neatly presented for you.

Components of a view

The tabular display has different components that show you the complete experiment history as well as enable to you to generate plots, compare experiments, run new experiments, etc. The major components of a view are:

• Git history - represents your experimentation history.
• Display preferences - show/hide branches, commits and columns, and re-arrange your view.
• Buttons to visualize, compare, and run experiments.

Git history

The branches and commits in your Git repository are displayed along with the corresponding metrics, files and hyperparameters.

Display preferences

The table contains buttons to specify filters and other preferences regarding which commits and columns to display.

Filters:

You can filter the commits that you want to display by the following fields:

• Branch: The Git branch
• Tag: The Git tag
• Author: Author of the Git commit
• Metric: Values of different metrics. For instance, you can display only those experiments for which the value of avg_prec is greater than 0.9.
• Metric delta: Change in the value of the metric. For instance, you can use this filter to only display those experiments for which the value of avg_prec changed by more than 0.1 compared to the baseline experiment.
• Param: Values of different parameters
• File size: Size of the data, model and other files corresponding to your experiments
• File changed: Whether or not any given file changed in the experiment

Columns:

Select the columns you want to display and hide the rest.

Additionally, you can click and drag the columns in the table to rearrange them as per your preferences.

If your view is missing some required columns or includes columns that you do not want, refer to the following troubleshooting sections to understand why this may have happened.

• View does not contain the columns that I want
• View contains columns that I did not import

Hide commits:

The following functionality are available for you to hide irrelevant commits from your views in DVC Studio.

• DVC Studio auto-hides irrelevant commits: DVC Studio identifies commits where metrics, files and hyperparameters did not change and hides them automatically.
• Hide commits and branches manually: You can selectively hide commits and branches. This can be useful useful if there are commits that do not add much value in your view. To hide a commit or branch, click on the 3-dot menu next to the commit or branch name and click on Hide commit or Hide branch.
• Unhide commits: You can unhide commits as needed, so that you don't lose any experimentation history. To display all hidden commits, click on the Show hidden commits toggle (refer the above gif). This will display all hidden commits, with a hidden (closed eye) indicator. To unhide any commit, click on the 3-dot menu for that commit and click on Show commit.

Selected only:

Use this toggle switch to show/hide experiments that you have not selected.

Delta mode:

Toggle between absolute values and difference from the first row.

Save changes:

Save your filters or column display preferences so that these preferences remain intact even after you log out of DVC Studio and log back in later.

Visualize, compare and run experiments.

The table also contains buttons to visualize, compare and run experiments.

• Show plots: Show plots for the selected commits. When you click on this button, plots for the selected commits are displayed in the bottom half of your browser window.
• Compare: Compare different experiments side by side.
• Run: Run experiments by selecting any one commit. Refer here for details on how to run experiments.
• Trends: Generate trend charts to show metric evolution over time.