Using Wandb with Rllib / Tune

RLlib
1

Hi!

I am trying to use the WandbLoggerCallback on Goolgle Colab while training my RL agent, however it does not seem to work:

tune.run(my_train_fn,
         checkpoint_freq=100,
         checkpoint_at_end=True,
         reuse_actors=True,
         keep_checkpoints_num=5,
         name="experiment",
         resources_per_trial=resources,
         config=config,
         callbacks=[
             WandbLoggerCallback(
                 api_key="12341234", # here I did put my correct key in my real project.. I replaced it for the example here
                 project="my_project" # I've created a PRIVATE project with the same name on the wandb platform
             )
         ]
)

I do not get any errors and no data is updated on the wandb platform … Am I missing something?
(Version: ray==1.2.0)

(Not sure if this post belongs to the Ray Tune or Rllib category. Please move if necessary)

Thanks for your time and help!
Walt