W&B Quickstart

Install W&B to track, visualize, and manage machine learning experiments of any size.

Sign up and create an API key

To authenticate your machine with W&B, generate an API key from your user profile or at wandb.ai/authorize. Copy the API key and store it securely.

Install the wandb library and log in

• Command Line
• Python
• Python notebook

pip install wandb

import wandb

wandb.login()

!pip install wandb
import wandb
wandb.login()

Start a run and track hyperparameters

In your Python script or notebook, initialize a W&B run object with wandb.init(). Use a dictionary for the config parameter to specify hyperparameter names and values.

run = wandb.init(
    project="my-awesome-project",  # Specify your project
    config={                        # Track hyperparameters and metadata
        "learning_rate": 0.01,
        "epochs": 10,
    },
)

A run serves as the core element of W&B, used to track metrics, create logs, and more.

Assemble the components

This mock training script logs simulated accuracy and loss metrics to W&B:

# train.py
import wandb
import random

wandb.login()

epochs = 10
lr = 0.01

run = wandb.init(
    project="my-awesome-project",    # Specify your project
    config={                         # Track hyperparameters and metadata
        "learning_rate": lr,
        "epochs": epochs,
    },
)

offset = random.random() / 5
print(f"lr: {lr}")

# Simulate a training run
for epoch in range(2, epochs):
    acc = 1 - 2**-epoch - random.random() / epoch - offset
    loss = 2**-epoch + random.random() / epoch + offset
    print(f"epoch={epoch}, accuracy={acc}, loss={loss}")
    wandb.log({"accuracy": acc, "loss": loss})

# run.log_code()

Visit wandb.ai/home to view recorded metrics such as accuracy and loss and how they changed during each training step. The following image shows the loss and accuracy tracked from each run. Each run object appears in the Runs column with generated names.

Next steps

Explore more features of the W&B ecosystem:

1. Read the W&B Integration tutorials that combine W&B with frameworks like PyTorch, libraries like Hugging Face, and services like SageMaker.
2. Organize runs, automate visualizations, summarize findings, and share updates with collaborators using W&B Reports.
3. Create W&B Artifacts to track datasets, models, dependencies, and results throughout your machine learning pipeline.
4. Automate hyperparameter searches and optimize models with W&B Sweeps.
5. Analyze runs, visualize model predictions, and share insights on a central dashboard.
6. Visit W&B AI Academy to learn about LLMs, MLOps, and W&B Models through hands-on courses.
7. Visit the official W&B Weave documentation to learn how to track track, experiment with, evaluate, deploy, and improve your LLM-based applications using Weave.