Experiment tracking
for deep learning

Record and visualize every detail of your research, collaborate easily, advance the state of the art

Sign UpRequest Demo

Record Everything

We store all your inputs and outputs in one place, making it easy to track and reproduce models. We even save your most recent git SHA for you.

Visualize Training

Default plots let you quickly see trends and compare runs in real time. Log a custom visualization and we'll save it with your run.

Simple and Powerful

Instrumenting your existing code takes less than five minutes and scales easily up to massive distributed training runs.

"W&B is a key piece of our fast-paced, cutting-edge, large-scale research workflow: great flexibility, performance, and user experience."

Adrien Gaidon
Toyota Research Institute

"W&B allows us to scale up insights from a single researcher to the entire team and from a single machine to thousands."

Wojciech Zaremba
Cofounder of OpenAI

Read the full interview with Peter Welinder from OpenAI ➞

Quick Integration

Add just a few lines of code to your script to get the power of W&B logging. We're framework agnostic.

import torch
import torch.nn as nn
from torchvision import datasets
import wandb

wandb.init()
...

Log Anything

Easily log metrics from your script to visualize results in real time as your model trains.

wandb.log({"acc": accuracy, "val_acc": val_accuracy})

Visualize Outputs

Peek under the hood and see what your model is producing at each time step.

wandb.log({"examples": fashion-mnist})

Massive Scale

Our service was built to handle thousands of parallel runs. Every experiment is saved for quick comparison in the run table. Tag, filter, sort, and group runs easily.

Compare Models

Visualize the influence of different architectures and hyperparameters on the output metrics you care about.

Example Projects

Keras MNIST Example
This sample project is classifying images of handwritten digits from the MNIST dataset.
iNaturalist Fine-Tuning
Stacey walks you through fine-tuning a Keras model to classify images of plants and animals.
Pytorch Image Colorizer
Boris trained a model to colorize black and white images, and he describes his process in this report.

Never lose your progress again.

Create a free account