Anonymous Mode

Log and visualize data without a W&B account

Are you writing a paper to submit to a conference? Use anonymous mode to let anyone run your code and get a Weights & Biases dashboard without creating an account.
wandb.init(anonymous="allow")
Example usage
​Try the example notebook to see how anonymous mode works in action.
import wandb
import random
​
wandb.init(project="anon-demo", 
           anonymous="allow",
           config={
               "learning_rate": 0.1,
               "batch_size": 128,
           })
​
for step in range(30):
  wandb.log({
      "acc": random.random(),
      "loss": random.random()
  })

System Metrics

Java Library [Beta]

Last updated 4 weeks ago