Are there any alterantives to tensorboard that is supported by Ray or rllib?
Ideally one that requires minimal changes to code, maybe just pointing it to to ~/ray_results is sufficient, like Tensorboard.
Examples: Neptune, MLflow, Comet, Sacred, Weights & Bases.
Neptune: it's definitely possible but I don't think there are any open-source examples
MLflow: https://github.com/ray-project/ray/blob/master/python/ray/tune/examples/mlflow_example.py
Comet: not sure, but probably the same case as Neptune
Sacred: this might work https://gitlab.com/SumNeuron/extune
Weights & Biases: https://docs.wandb.com/library/integrations/ray-tune
If you go with Neptune or Comet, I'd recommend you look at WandB's custom logger鈥攊t's really just a shim API from Ray to WandB.
hey @vertangelx and @sumanthratna,
Here is Kamil from neptune.ai. We are now working on integration with tune.
Do you have any specific needs regarding it? - it's really great time to share a bit about your needs :)
Integration will work as a tune logger (similar to MLFlow), so you can use it with minimal code changes, like any other Logger. It will naturally work with rllib.
Hi, I'm a bot from the Ray team :)
To help human contributors to focus on more relevant issues, I will automatically add the stale label to issues that have had no activity for more than 4 months.
If there is no further activity in the 14 days, the issue will be closed!
You can always ask for help on our discussion forum or Ray's public slack channel.
Hi, @kamil-kaczmarek, we are using Neptune with RLLIB in our projects currently. Are there any updates on the integration with tune? We are really looking forward to this feature! Especially now wandb has been supported ;) Thanks!
Hi @mathpluscode,
We are working on it in Neptune. I hope that we will be able to release it in the upcoming weeks. Very happy to hear that you use Neptune with RLlib!