The W&B Reports and Workspaces API, accessible at wandb_workspaces
, allows you to create reports, which can be published on the web to share findings, and well as customize a workspace where training and fine-tuning work was done.
W&B Report and Workspace API is in Public Preview.
Installation and setup
Sign up and create an API key
To authenticate your machine with W&B, you must first generate an API key at https://wandb.ai/authorize.
Install and import packages
Install the W&B Report and Workspaces library.
pip install wandb-workspaces
Create a report
To create a report, specify your team’s entity and provide a name for your report. Replace enclosed text with your values:
import wandb_workspaces.reports.v2 as wr
# Create
report = wr.Report(
entity="<team_entity>",
project="<project_name>",
title='Quickstart Report',
description="That was easy!"
)
# Save report
report.save()
Next, add blocks and panels to your report. For example, the following code creates a report with a table of contents, header, and a paragraph:
report.blocks = [
wr.TableOfContents(),
wr.H1("Text and images example"),
wr.P("Lorem ipsum dolor sit amet."),
]
report.save()
See the Reports API Quickstart Google Colab for an end to end example.
Create a workspace
The following code shows how to create a workspace with a section containing three panels: a line plot, a bar plot, and a scalar chart. Replace enclosed text with your values:
# How to import
import wandb_workspaces.workspaces as ws
# Create a workspace
ws.Workspace(
entity="<team_entity>", # entity that owns the workspace
project="<project_name>", # project that the workspace is associated with
sections=[
ws.Section(
name="<Validation Metrics>",
panels=[
wr.LinePlot(x="Step", y=["<val_loss>"]),
wr.BarPlot(metrics=["<val_accuracy>"]),
wr.ScalarChart(metric="<f1_score>", groupby_aggfunc="<mean>"),
],
is_open=True,
),
],
)
workspace.save()
See the Workspace API Quickstart Google Colab for an end to end example.