How to beginner · 3 min read

wandb dashboard explained

Q: wandb dashboard explained

The wandb dashboard is a web interface that visualizes machine learning experiment metrics, logs, and artifacts in real time. It helps developers track training progress, compare runs, and analyze model performance efficiently.

Quick answer

The wandb dashboard is a web interface that visualizes machine learning experiment metrics, logs, and artifacts in real time. It helps developers track training progress, compare runs, and analyze model performance efficiently.

PREREQUISITES

Python 3.8+
pip install wandb
Weights & Biases account (free tier available)
Basic knowledge of machine learning training loops

Setup

Install the wandb Python package and log in to your Weights & Biases account to enable experiment tracking.

bash

pip install wandb
wandb login

output

wandb: Currently logged in as: your_username (use `wandb login --relogin` to force relogin)

Step by step

Integrate wandb into your training script to log metrics and visualize them on the dashboard.

python

import wandb

# Initialize a new run
wandb.init(project="my-ml-project", entity="your_username")

for epoch in range(5):
    # Simulate training metric
    accuracy = 0.8 + epoch * 0.04
    loss = 0.5 / (epoch + 1)

    # Log metrics to wandb
    wandb.log({"epoch": epoch, "accuracy": accuracy, "loss": loss})

print("Training complete. Check your wandb dashboard for live updates.")

output

Training complete. Check your wandb dashboard for live updates.

Common variations

You can customize the dashboard by logging additional data types such as images, model checkpoints, and hyperparameters. Use wandb.watch() to automatically log gradients and model topology.

python

import wandb
import torch
import torch.nn as nn

wandb.init(project="my-ml-project", entity="your_username")

model = nn.Linear(10, 2)
wandb.watch(model, log="all")  # Logs gradients and parameters

# Log hyperparameters
wandb.config.update({"learning_rate": 0.001, "batch_size": 32})

# Log an image example
wandb.log({"sample_image": wandb.Image(torch.randn(3, 64, 64))})

Troubleshooting

If metrics do not appear on the dashboard, ensure you called wandb.init() before logging and that your internet connection is active.
Use wandb login to authenticate if you see authorization errors.
Check the wandb run directory for logs if runs fail to sync.

✅

Key Takeaways

Use wandb.init() to start tracking experiments and wandb.log() to send metrics to the dashboard.
The dashboard provides real-time visualization of metrics, enabling easy comparison and analysis of multiple runs.
Log additional data like images, hyperparameters, and model gradients to enrich experiment insights.
Ensure proper authentication with wandb login and stable internet for syncing data.
The dashboard is accessible via your Weights & Biases web account and supports collaboration.

Verified 2026-04

Verify ↗