A CLI and library for interacting with the Weights and Biases API.
Project description
Weights and Biases
The Weights and Biases client is an open source library, CLI (wandb), and local web application for organizing and analyzing your machine learning experiments. Think of it as a framework-agnostic lightweight TensorBoard that persists additional information such as the state of your code, system metrics, and configuration parameters.
Local Features
- Store config parameters used in a training run
- Associate version control with your training runs
- Search, compare, and visualize training runs
- Analyze system usage metrics alongside runs
Cloud Features
- Collaborate with team members
- Run parameter sweeps
- Persist runs forever
Quickstart
pip install wandb
In your training script:
import wandb
from wandb.keras import WandbCallback
# Your custom arguments defined here
args = ...
run = wandb.init(config=args)
run.config["more"] = "custom"
def training_loop():
while True:
# Do some machine learning
epoch, loss, val_loss = ...
# Framework agnostic / custom metrics
run.history.add({"epoch": epoch, "loss": loss, "val_loss": val_loss})
# Keras metrics
model.fit(..., callbacks=[WandbCallback()])
Running your training script will save data in a directory named wandb relative to your training script. To view your runs, call wandb board
from the same directory as your training script.
Cloud Usage
Signup for an account, then run wandb init
from the directory with your training script. You can checkin wandb/settings to version control to enable other users on your team to share experiments. Run your script with wandb run my_script.py
and all metadata will be synced to the cloud.
Detailed Usage
Framework specific and detailed usage can be found in our documentation.
Development
See https://github.com/wandb/client/blob/master/DEVELOPMENT.md
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for wandb-0.6.2-py2.py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 49f67da8802c70cbe695ff7267b60066f7552c869cc22c6cff45740d3d5486b2 |
|
MD5 | ae6179ac6d30436d84ee75b43743abff |
|
BLAKE2b-256 | e647cf72a8e6cacbc7aed740d9c68ba3c84f8cc8d2ed358a21f2bd26920b47b3 |