wandb.init() returns a run object. You can also access the run object anywhere in your code by importing wandb and calling
wandb.run (as long as wandb.init() has already been called).
You should generally call
wandb.init() once at the start of your training script.
In a Jupyter Notebook, calling
wandb.init() will create a new run.
wandb.init() accepts a few keyword arguments:
- name — A display name for this run
- notes — A multiline string description associated with the run
- config — a dictionary-like object to set as initial config
- project — the name of the project to which this run will belong
- tags — a list of strings to associate with this run as tags
- dir — the path to a directory where artifacts will be written (default: ./wandb)
- entity — the team posting this run (default: your username or your default team)
- job_type — the type of job you are logging, e.g. eval, worker, ps (default: training)
- group — a string by which to group other runs; see Grouping
- sync_tensorboard — A boolean indicating whether or not copy all tensorboard logs wandb. see Tensorboard (default: False)
- reinit — whether to allow multiple calls to wandb.init in the same process (default: False)
- id — A unique id for this run primarily used for resuming see Resuming, must be globally unique within a project
- resume — if set to True, the run auto resumes; can also be a unique string for manual resuming; see Resuming (default: False)