How do I update run config, tags, and notes via the W&B API?

After a run finishes you can edit its config values, display name, tags, and notes using the Public API without re-running the experiment. Updating config values

import wandb

api = wandb.Api()
run = api.run("my-entity/my-project/run_id_here")

# Add or overwrite config keys
run.config["post_hoc_label"] = "baseline-v2"
run.config["dataset_version"] = 3
run.update()

run.update() persists all pending changes (config, name, tags, notes) in a single API call. Renaming a run

run = api.run("my-entity/my-project/run_id_here")
run.name = "resnet50-lr1e-3-aug-v2"
run.update()

Adding and removing tags

run = api.run("my-entity/my-project/run_id_here")

# Add tags
run.tags.append("production-ready")
run.tags.append("reviewed")
run.update()

# Remove a specific tag
run.tags = [t for t in run.tags if t != "wip"]
run.update()

Updating run notes Notes appear in the run detail page and support Markdown:

run = api.run("my-entity/my-project/run_id_here")
run.notes = "Final baseline. Trained on v3 dataset with Aug-v2. Val accuracy: 92.4%."
run.update()

Bulk-updating many runs To apply changes across a set of runs (e.g., tag all runs from a sweep as reviewed):

runs = api.runs(
    "my-entity/my-project",
    filters={"sweep": "sweep_id_here", "state": "finished"}
)

for run in runs:
    run.tags.append("sweep-reviewed")
    run.update()

Iterate carefully — each run.update() is a separate API call. For very large sets, add a short sleep between calls to avoid hitting rate limits. What you cannot change after a run finishes You cannot add new logged metrics (summary values or history) to a finished run via the API. The summary object is readable but not writable through the Public API. If you need to attach computed post-hoc metrics, use a new run with wandb.init(job_type="evaluation") and log there, then link it to the original via tags or artifact lineage.

Runs Experiments API

Documentation Index