Pin and compare runs

• Pinned runs: Pin runs from any project in your workspace to keep them visible at the top of the runs list. You can pin up to 20 runs in a workspace, including runs from other projects.
• Baseline run: Specify a baseline run as your reference point for comparisons. The baseline run is always visible in the workspace and at the top of the runs list. In the runs table, summary metric deltas show how each run compares to the baseline. In line plots, the baseline appears with visually distinct styling to help with comparison.

• Comparing new experiments against your production model.
• Comparing runs across projects.
• Creating a baseline for your experiments to evaluate how new runs perform against it.
• Tracking multiple candidate models during experimentation.
• Evaluating whether new runs improve on your best results.

​

Pin runs

1. Navigate to your workspace.
2. In the run selector or runs table, find the run you want to pin.
3. Click the action () menu, then select Pin run.

​

Compare runs across projects

1. Navigate to your project’s workspace.
2. Click on the Select runs from another project button (small square box with rounded corners and a diagonal arrow) at the top of your workspace’s run selector or the runs table.

3. In the modal, select a project from the Source project dropdown.
4. Search the source project’s runs to find the run you want to compare. You can search by run name or unique ID.
5. Select the checkbox next to each run you want to compare, then click Pin runs.

​

Manage and compare a baseline run

​

Set a baseline run

1. Navigate to your workspace.
2. In the run selector or runs table, find the run you want to use as your baseline.
3. Click the action () menu, then select Set as baseline.

​

Change the baseline run

1. Navigate to your workspace.
2. In the run selector or runs table, find the run you want to use as your new baseline.
3. Click the action () menu, then select Replace baseline.
   If the menu item is inactive, ensure that you have at least one pinning slot available. If necessary, unpin a pinned run by clicking the circular pin icon next to a pinned run.
4. The new run becomes the baseline, and W&B automatically pins the previous baseline so you can find it. Optionally, unpin it by clicking its pin icon.

​

Remove the baseline designation

1. Navigate to your workspace.
2. In the run selector or runs table, find the current baseline run.
3. Click the action () menu, then select Remove baseline.
   If the menu item is inactive, ensure that you have at least one pinning slot available. If necessary, unpin a pinned run by clicking the circular pin icon next to a pinned run.
4. W&B automatically pins the previous baseline so you can find it. Optionally, unpin it by clicking its pin icon.

​

Compare runs to the baseline

• Hover over a part of the plot to display a tooltip with values for all visible runs, including the baseline run and pinned runs.
  
• Hover over the baseline run’s legend label to display the line prominently. It appears as a heavy dashed line. Lines for other visible runs appear with reduced saturation.
  
• Hover over another run’s legend label to display that run’s line prominently and compare it with the baseline, which appears as a heavy dashed line. Lines for other visible runs appear with reduced saturation.
  

​

Summary metric deltas

• If the other run outperforms (is directionally better than) the baseline, the delta appears as dark red text with a light red background.
• If the other run underperforms (is directionally worse than) the baseline, the delta appears as dark teal text with a light teal background.

1. In the runs table, hover over the column heading for the metric.
2. Click the action () menu that appears.
3. Set Metric directionality to Higher values are best or Lower values are best.

​

Hide summary metric deltas in a workspace

1. In the workspace, click Settings.
2. In the drawer that appears, click Runs.
3. In the Baseline tab, toggle Show value deltas in the runs table.
4. Close the workspace settings drawer.

​

Use cases

• Track production models: Ensure that new models meet your quality bar before deployment.
  1. Set your production model as the baseline.
  2. Compare all experiments against your deployed model to identify candidates that outperform production.
• Compare hyperparameter experiments: Evaluate hyperparameter sweeps or manual experiments against your best-known configuration.
  1. Set your best-known configuration as the baseline.
  2. Pin promising candidates as you discover them.
  3. Use line plots to compare runs against the baseline.
  4. Continue to update the baseline as you find better configurations.

​

Example: Compare runs from hyperparameter tuning experiments

import wandb
import random
import math

def train_model(learning_rate, batch_size, run_name, tags=None):
    """Simulate training a model with given hyperparameters."""
    config = {
        "learning_rate": learning_rate,
        "batch_size": batch_size,
        "optimizer": "adam",
        "architecture": "resnet50"
    }
    
    with wandb.init(
      # Replace with your team and project name
        project="hyperparameter-tuning",
        entity="your-team-name", # Replace with your team name
        name=run_name,
        config=config,
        tags=tags or []
    ) as run:
        # Simulate training loop
        for epoch in range(50):
            # Simulated metrics
            accuracy = 0.6 + 0.3 * (1 - math.exp(-learning_rate * epoch / 10))
            loss = 1.0 * math.exp(-learning_rate * epoch / 10)
            
            run.log({
                "epoch": epoch,
                "accuracy": accuracy,
                "loss": loss
            })

# Create baseline run with standard configuration
train_model(
    learning_rate=0.001,
    batch_size=64,
    run_name="baseline-config",
    tags=["baseline", "production"]
)

# Experiment with different learning rates
train_model(
    learning_rate=0.003,
    batch_size=64,
    run_name="lr-experiment-0.003",
    tags=["experiment"]
)

train_model(
    learning_rate=0.0001,
    batch_size=64,
    run_name="lr-experiment-0.0001",
    tags=["experiment"]
)

1. Navigate to your W&B project.
2. Select the Runs tab.
3. Filter for the run that has the baseline tag and set it as your baseline run.
4. Pin baseline-config to make the run visible at the top of the runs table.
5. Compare the experiment runs against the baseline.
   • Select the vertical ellipsis menu next to the accuracy column, then set Metric directionality to Higher values are best. The delta for each run’s accuracy is now shown in the runs table.
   • Select the vertical ellipsis menu next to the loss column, then set Metric directionality to Lower values are best. The delta for each run’s loss is now shown in the runs table.
   
6. Use the line plots to visually compare the runs against the baseline. The baseline run appears with a bold dashed line, and you can hover over the plot to see how each run compares at a given epoch. For example, the following image shows how the runs lr-experiment-0.003 and lr-experiment-0.0001 compare with the baseline run baseline-config.
   

​

Limitations

• Grouping: When viewing runs in the run selector or runs table, if runs are grouped by a column, pinned and baseline runs aren’t visually distinct from other runs.
• Reports: In a run set in a W&B Report, pinned and baseline runs aren’t visually distinct from other runs.
• Workspace view only: The baseline doesn’t appear when viewing a single run’s workspace.
• Line plots only: Baseline comparison is available only for line plots, not for other panels such as bar plots or media panels.