Weights & Biases
App · 18 actions
Work with your Weights & Biases experiments: find and tag runs, read their metrics, record results, promote model versions with aliases, and list reports and sweeps. Start a workflow when a run finishes, fails or crosses a metric threshold, or when a new model version lands.
Actions
| Action | What it does |
|---|---|
| Add artifact aliases | Put aliases such as production on a version of a W&B artifact. |
| Create project | Create a W&B project. |
| Delete run | Delete a W&B run, and optionally the artifacts it logged. |
| Get artifact version | Get one version of a W&B artifact by its version or alias, with its metadata. |
| Get my account | Get the W&B account the API key belongs to: username, email, default team and every team you belong to. |
| Get project | Get a W&B project, with how many runs it has. |
| Get run | Get a W&B run with its config, latest metrics (summary) and system metrics. |
| Get run metric history | Get the values a W&B run logged over time, such as its loss at each step, sampled down to the number of rows asked for. |
| Get sweep | Get a W&B hyperparameter sweep: its state, settings, run count and best result. |
| List artifact versions | List the versions of a W&B artifact, with their aliases. |
| List artifacts | List the artifacts (models, datasets and other file sets) in a W&B project. |
| List projects | List the projects in a W&B team or your personal account. |
| List reports | List the reports in a W&B project, with a link to each. |
| List runs | List the runs in a W&B project, newest first, optionally only those in a state, with a tag, in a group or started after a time. |
| List sweeps | List the hyperparameter sweeps in a W&B project, with each one's state and best result. |
| Remove artifact aliases | Take aliases off a version of a W&B artifact. |
| Set run summary metrics | Add or change values in a W&B run's summary, such as a review score or a test result. |
| Update run | Rename a W&B run, change its notes or group, or add or remove tags. |
Add artifact aliases
Put aliases such as production on a version of a W&B artifact. An alias is on one version at a time, so it moves from any other.
Inputs
- Team
entitytextRequired - The W&B team, or your own username, that owns the project.
- Project
projecttextRequired - The W&B project. Pick one, or give its name.
- Artifact version
artifacttextRequired - The artifact's name and version or alias, such as my-model:v3 or my-model:production. A name alone means its latest version.
- Aliases
aliaseslistRequired - One or more aliases, such as production or staging.
Outputs
What the step returns
artifactobject- The version as it is now
changedtrue/false- False if nothing needed changing, so nothing was sent
Example output
{
"artifact": {
"aliases": [
"latest",
"production"
],
"collection": "resnet50",
"created_at": "2026-11-05T09:31:00",
"description": "",
"digest": "a1b2c3d4",
"entity": "acme-ml",
"file_count": 2,
"id": "QXJ0aWZhY3Q6NjQ1MjE=",
"metadata": {
"framework": "pytorch"
},
"name": "resnet50:v3",
"project": "image-classifier",
"size": 102400000,
"state": "COMMITTED",
"type": "model",
"updated_at": "2026-11-05T09:31:00",
"version": "v3",
"version_index": 3
},
"changed": true
}Create project
Create a W&B project. If one with that name already exists it is returned as it is, unchanged.
Inputs
- Team
entitytextRequired - The W&B team, or your own username, that owns the project.
- Project name
projecttextRequired - The new project's name: letters, digits, '-', '_' and '.', such as image-classifier.
- Description
descriptiontext - What the project is for.
Outputs
What the step returns
createdtrue/false- False if it already existed and was left as it was
projectobject- The project
Example output
{
"created": true,
"project": {
"created_at": "2026-09-01T10:00:00",
"description": "ResNet experiments",
"entity": "acme-ml",
"id": "UHJvamVjdDp2MTpteS1wcm9qZWN0OmFjbWU=",
"name": "image-classifier",
"url": "https://wandb.ai/acme-ml/image-classifier/workspace"
}
}Delete run
Delete a W&B run, and optionally the artifacts it logged. This cannot be undone.
Inputs
- Team
entitytextRequired - The W&B team, or your own username, that owns the project.
- Project
projecttextRequired - The W&B project. Pick one, or give its name.
- Run
run_idtextRequired - The run. Pick one of the project's latest runs, or give its ID or link from an earlier step.
- Also delete its artifacts
delete_artifactstrue/false - Also delete the artifacts the run logged.Default:
false.
Outputs
What the step returns
artifacts_deletedtrue/false- Whether its artifacts were deleted too
deletedtrue/false- True once it is deleted
display_nametext- Its name
run_idtext- The deleted run's ID
Example output
{
"artifacts_deleted": false,
"deleted": true,
"display_name": "eager-fox-12",
"run_id": "1mbku38n"
}Get artifact version
Get one version of a W&B artifact by its version or alias, with its metadata.
Inputs
- Team
entitytextRequired - The W&B team, or your own username, that owns the project.
- Project
projecttextRequired - The W&B project. Pick one, or give its name.
- Artifact version
artifacttextRequired - The artifact's name and version or alias, such as my-model:v3 or my-model:production. A name alone means its latest version.
Outputs
What the step returns
artifactobject- The artifact version
Example output
{
"artifact": {
"aliases": [
"latest",
"production"
],
"collection": "resnet50",
"created_at": "2026-11-05T09:31:00",
"description": "",
"digest": "a1b2c3d4",
"entity": "acme-ml",
"file_count": 2,
"id": "QXJ0aWZhY3Q6NjQ1MjE=",
"metadata": {
"framework": "pytorch"
},
"name": "resnet50:v3",
"project": "image-classifier",
"size": 102400000,
"state": "COMMITTED",
"type": "model",
"updated_at": "2026-11-05T09:31:00",
"version": "v3",
"version_index": 3
}
}Get my account
Get the W&B account the API key belongs to: username, email, default team and every team you belong to.
Inputs
None.
Outputs
What the step returns
userobject- Your W&B account
Example output
{
"user": {
"admin": false,
"default_entity": "amelie",
"email": "amelie@example.com",
"id": "VXNlcjoxMjM0NQ==",
"name": "Amelie Pangolin",
"teams": [
"amelie",
"acme-ml"
],
"username": "amelie"
}
}Get project
Get a W&B project, with how many runs it has.
Inputs
- Team
entitytextRequired - The W&B team, or your own username, that owns the project.
- Project
projecttextRequired - The W&B project. Pick one, or give its name.
Outputs
What the step returns
projectobject- The project
Example output
{
"project": {
"created_at": "2026-09-01T10:00:00",
"description": "ResNet experiments",
"entity": "acme-ml",
"id": "UHJvamVjdDp2MTpteS1wcm9qZWN0OmFjbWU=",
"name": "image-classifier",
"read_only": false,
"run_count": 42,
"url": "https://wandb.ai/acme-ml/image-classifier/workspace"
}
}Get run
Get a W&B run with its config, latest metrics (summary) and system metrics.
Inputs
- Team
entitytextRequired - The W&B team, or your own username, that owns the project.
- Project
projecttextRequired - The W&B project. Pick one, or give its name.
- Run
run_idtextRequired - The run. Pick one of the project's latest runs, or give its ID or link from an earlier step.
Outputs
What the step returns
runobject- The run
Example output
{
"run": {
"commit": "3f2c1ab",
"config": {
"epochs": 20,
"learning_rate": 0.001
},
"created_at": "2026-11-05T08:00:00",
"display_name": "eager-fox-12",
"entity": "acme-ml",
"group": "resnet50",
"heartbeat_at": "2026-11-05T09:30:00",
"history_steps": 1200,
"job_type": "train",
"notes": "",
"project": "image-classifier",
"run_id": "1mbku38n",
"state": "finished",
"storage_id": "UnVuOnYxOjFtYmt1MzhuOmltYWdlLWNsYXNzaWZpZXI6YWNtZS1tbA==",
"summary": {
"_step": 1199,
"val_accuracy": 0.93,
"val_loss": 0.21
},
"sweep_id": null,
"system_metrics": {},
"tags": [
"baseline"
],
"url": "https://wandb.ai/acme-ml/image-classifier/runs/1mbku38n",
"user": "amelie"
}
}Get run metric history
Get the values a W&B run logged over time, such as its loss at each step, sampled down to the number of rows asked for.
Inputs
- Team
entitytextRequired - The W&B team, or your own username, that owns the project.
- Project
projecttextRequired - The W&B project. Pick one, or give its name.
- Run
run_idtextRequired - The run. Pick one of the project's latest runs, or give its ID or link from an earlier step.
- Metrics
metricslist - Only these metrics, such as loss and accuracy. Leave blank for everything the run logged.
- Rows
samplesnumber - How many rows at most, up to 5000. Defaults to 500.From 1 to 5000.
Outputs
What the step returns
countnumber- How many rows were returned
rowslist- One row per logged step, with _step and each metric
run_idtext- The run's ID
Example output
{
"count": 2,
"rows": [
{
"_step": 0,
"loss": 2.3
},
{
"_step": 100,
"loss": 0.9
}
],
"run_id": "1mbku38n"
}Get sweep
Get a W&B hyperparameter sweep: its state, settings, run count and best result.
Inputs
- Team
entitytextRequired - The W&B team, or your own username, that owns the project.
- Project
projecttextRequired - The W&B project. Pick one, or give its name.
- Sweep
sweep_idtextRequired - The sweep. Pick one, or give its ID or link.
Outputs
What the step returns
sweepobject- The sweep
Example output
{
"sweep": {
"best_loss": 0.19,
"config": {
"method": "bayes",
"metric": {
"goal": "minimize",
"name": "val_loss"
}
},
"created_at": "2026-11-04T12:00:00",
"description": "",
"display_name": "lr search",
"entity": "acme-ml",
"method": "bayes",
"project": "image-classifier",
"run_count": 14,
"run_count_expected": 30,
"state": "RUNNING",
"storage_id": "U3dlZXA6djE6eDdrMnA5cWE=",
"sweep_id": "x7k2p9qa",
"updated_at": "2026-11-05T09:00:00",
"url": "https://wandb.ai/acme-ml/image-classifier/sweeps/x7k2p9qa"
}
}List artifact versions
List the versions of a W&B artifact, with their aliases.
Inputs
- Team
entitytextRequired - The W&B team, or your own username, that owns the project.
- Project
projecttextRequired - The W&B project. Pick one, or give its name.
- Artifact
collectiontextRequired - The artifact, such as my-model.
- Artifact type
artifact_typetext - The artifact's type. Leave blank to look it up.
- Limit
limitnumber - How many to return, at most 500. Defaults to 50.From 1 to 500.
Outputs
What the step returns
countnumber- How many were returned
versionslist- The versions
Example output
{
"count": 1,
"versions": [
{
"aliases": [
"latest",
"production"
],
"collection": "resnet50",
"created_at": "2026-11-05T09:31:00",
"description": "",
"digest": "a1b2c3d4",
"entity": "acme-ml",
"file_count": 2,
"id": "QXJ0aWZhY3Q6NjQ1MjE=",
"metadata": {
"framework": "pytorch"
},
"name": "resnet50:v3",
"project": "image-classifier",
"size": 102400000,
"state": "COMMITTED",
"type": "model",
"updated_at": "2026-11-05T09:31:00",
"version": "v3",
"version_index": 3
}
]
}List artifacts
List the artifacts (models, datasets and other file sets) in a W&B project.
Inputs
- Team
entitytextRequired - The W&B team, or your own username, that owns the project.
- Project
projecttextRequired - The W&B project. Pick one, or give its name.
- Artifact type
artifact_typetext - Only artifacts of this type, such as model or dataset.
- Limit
limitnumber - How many to return, at most 500. Defaults to 50.From 1 to 500.
Outputs
What the step returns
collectionslist- The artifacts
countnumber- How many were returned
Example output
{
"collections": [
{
"created_at": "2026-10-01T10:00:00",
"description": "",
"id": "QXJ0aWZhY3RDb2xsZWN0aW9uOjEyMw==",
"name": "resnet50",
"type": "model",
"updated_at": "2026-11-05T09:31:00"
}
],
"count": 1
}List projects
List the projects in a W&B team or your personal account.
Inputs
- Team
entitytextRequired - The W&B team, or your own username, that owns the project.
- Limit
limitnumber - How many to return, at most 500. Defaults to 50.From 1 to 500.
Outputs
What the step returns
countnumber- How many were returned
entitytext- The team or user listed
projectslist- The projects
Example output
{
"count": 1,
"entity": "acme-ml",
"projects": [
{
"created_at": "2026-09-01T10:00:00",
"description": "ResNet experiments",
"entity": "acme-ml",
"id": "UHJvamVjdDp2MTpteS1wcm9qZWN0OmFjbWU=",
"name": "image-classifier",
"url": "https://wandb.ai/acme-ml/image-classifier/workspace"
}
]
}List reports
List the reports in a W&B project, with a link to each.
Inputs
- Team
entitytextRequired - The W&B team, or your own username, that owns the project.
- Project
projecttextRequired - The W&B project. Pick one, or give its name.
- Limit
limitnumber - How many to return, at most 500. Defaults to 50.From 1 to 500.
Outputs
What the step returns
countnumber- How many were returned
reportslist- The reports
Example output
{
"count": 1,
"reports": [
{
"author": "amelie",
"created_at": "2026-11-02T10:00:00",
"description": "",
"entity": "acme-ml",
"id": "Vmlldzo1MjM0NTY3",
"project": "image-classifier",
"title": "Weekly results",
"updated_at": "2026-11-05T10:00:00",
"url": "https://wandb.ai/acme-ml/image-classifier/reports/Weekly-results--Vmlldzo1MjM0NTY3"
}
]
}List runs
List the runs in a W&B project, newest first, optionally only those in a state, with a tag, in a group or started after a time.
Inputs
- Team
entitytextRequired - The W&B team, or your own username, that owns the project.
- Project
projecttextRequired - The W&B project. Pick one, or give its name.
- State
statetext - Only runs in this state.One of:
running,finished,failed,crashed,killed,pending. - Tags
tagslist - Only runs with any of these tags.
- Group
grouptext - Only runs in this group.
- Started after
created_afterdate and time - Only runs started after this time.
- Order
ordertext - Which runs come first. Defaults to newest first.One of:
-created_at(Newest first (default)),+created_at(Oldest first),-heartbeat_at(Most recently active first). Default:-created_at. - Include summary metrics
include_metricstrue/false - Also return each run's latest metrics (its summary).Default:
false. - Limit
limitnumber - How many to return, at most 500. Defaults to 50.From 1 to 500.
Outputs
What the step returns
countnumber- How many were returned
runslist- The runs
Example output
{
"count": 1,
"runs": [
{
"commit": "3f2c1ab",
"created_at": "2026-11-05T08:00:00",
"display_name": "eager-fox-12",
"entity": "acme-ml",
"group": "resnet50",
"heartbeat_at": "2026-11-05T09:30:00",
"history_steps": 1200,
"job_type": "train",
"notes": "",
"project": "image-classifier",
"run_id": "1mbku38n",
"state": "finished",
"storage_id": "UnVuOnYxOjFtYmt1MzhuOmltYWdlLWNsYXNzaWZpZXI6YWNtZS1tbA==",
"summary": {},
"sweep_id": null,
"tags": [
"baseline"
],
"url": "https://wandb.ai/acme-ml/image-classifier/runs/1mbku38n",
"user": "amelie"
}
]
}List sweeps
List the hyperparameter sweeps in a W&B project, with each one's state and best result.
Inputs
- Team
entitytextRequired - The W&B team, or your own username, that owns the project.
- Project
projecttextRequired - The W&B project. Pick one, or give its name.
- Limit
limitnumber - How many to return, at most 500. Defaults to 50.From 1 to 500.
Outputs
What the step returns
countnumber- How many were returned
sweepslist- The sweeps
Example output
{
"count": 1,
"sweeps": [
{
"best_loss": 0.19,
"config": {
"method": "bayes",
"metric": {
"goal": "minimize",
"name": "val_loss"
}
},
"created_at": "2026-11-04T12:00:00",
"description": "",
"display_name": "lr search",
"entity": "acme-ml",
"method": "bayes",
"project": "image-classifier",
"run_count": 14,
"run_count_expected": 30,
"state": "RUNNING",
"storage_id": "U3dlZXA6djE6eDdrMnA5cWE=",
"sweep_id": "x7k2p9qa",
"updated_at": "2026-11-05T09:00:00",
"url": "https://wandb.ai/acme-ml/image-classifier/sweeps/x7k2p9qa"
}
]
}Remove artifact aliases
Take aliases off a version of a W&B artifact.
Inputs
- Team
entitytextRequired - The W&B team, or your own username, that owns the project.
- Project
projecttextRequired - The W&B project. Pick one, or give its name.
- Artifact version
artifacttextRequired - The artifact's name and version or alias, such as my-model:v3 or my-model:production. A name alone means its latest version.
- Aliases
aliaseslistRequired - One or more aliases, such as production or staging.
Outputs
What the step returns
artifactobject- The version as it is now
changedtrue/false- False if nothing needed changing, so nothing was sent
Example output
{
"artifact": {
"aliases": [
"latest"
],
"collection": "resnet50",
"created_at": "2026-11-05T09:31:00",
"description": "",
"digest": "a1b2c3d4",
"entity": "acme-ml",
"file_count": 2,
"id": "QXJ0aWZhY3Q6NjQ1MjE=",
"metadata": {
"framework": "pytorch"
},
"name": "resnet50:v3",
"project": "image-classifier",
"size": 102400000,
"state": "COMMITTED",
"type": "model",
"updated_at": "2026-11-05T09:31:00",
"version": "v3",
"version_index": 3
},
"changed": true
}Set run summary metrics
Add or change values in a W&B run's summary, such as a review score or a test result. The run's other summary metrics stay.
Inputs
- Team
entitytextRequired - The W&B team, or your own username, that owns the project.
- Project
projecttextRequired - The W&B project. Pick one, or give its name.
- Run
run_idtextRequired - The run. Pick one of the project's latest runs, or give its ID or link from an earlier step.
- Metrics
metricsobjectRequired - Each metric's name and value, such as test_accuracy: 0.91.
Outputs
What the step returns
changedtrue/false- False if nothing needed changing, so nothing was sent
runobject- The run as it is now
Example output
{
"changed": true,
"run": {
"commit": "3f2c1ab",
"config": {
"epochs": 20,
"learning_rate": 0.001
},
"created_at": "2026-11-05T08:00:00",
"display_name": "eager-fox-12",
"entity": "acme-ml",
"group": "resnet50",
"heartbeat_at": "2026-11-05T09:30:00",
"history_steps": 1200,
"job_type": "train",
"notes": "",
"project": "image-classifier",
"run_id": "1mbku38n",
"state": "finished",
"storage_id": "UnVuOnYxOjFtYmt1MzhuOmltYWdlLWNsYXNzaWZpZXI6YWNtZS1tbA==",
"summary": {
"_step": 1199,
"val_accuracy": 0.93,
"val_loss": 0.21
},
"sweep_id": null,
"system_metrics": {},
"tags": [
"baseline"
],
"url": "https://wandb.ai/acme-ml/image-classifier/runs/1mbku38n",
"user": "amelie"
}
}Update run
Rename a W&B run, change its notes or group, or add or remove tags. Fields left blank stay as they are.
Inputs
- Team
entitytextRequired - The W&B team, or your own username, that owns the project.
- Project
projecttextRequired - The W&B project. Pick one, or give its name.
- Run
run_idtextRequired - The run. Pick one of the project's latest runs, or give its ID or link from an earlier step.
- Display name
display_nametext - The run's new name in W&B.
- Notes
notestext - The run's new notes. They replace the current notes.
- Tags
tagslist - Tags to add or remove, such as best or reviewed.
- Tags change
tags_modetext - Add the tags, remove them, or replace all of them.One of:
add(Add to the run's tags (default)),remove(Remove from the run's tags),replace(Replace all of the run's tags). Default:add. - Group
grouptext - Move the run into this group.
Outputs
What the step returns
changedtrue/false- False if nothing needed changing, so nothing was sent
runobject- The run as it is now
Example output
{
"changed": true,
"run": {
"commit": "3f2c1ab",
"config": {
"epochs": 20,
"learning_rate": 0.001
},
"created_at": "2026-11-05T08:00:00",
"display_name": "eager-fox-12",
"entity": "acme-ml",
"group": "resnet50",
"heartbeat_at": "2026-11-05T09:30:00",
"history_steps": 1200,
"job_type": "train",
"notes": "",
"project": "image-classifier",
"run_id": "1mbku38n",
"state": "finished",
"storage_id": "UnVuOnYxOjFtYmt1MzhuOmltYWdlLWNsYXNzaWZpZXI6YWNtZS1tbA==",
"summary": {
"_step": 1199,
"val_accuracy": 0.93,
"val_loss": 0.21
},
"sweep_id": null,
"system_metrics": {},
"tags": [
"baseline",
"reviewed"
],
"url": "https://wandb.ai/acme-ml/image-classifier/runs/1mbku38n",
"user": "amelie"
}
}