[Core] Add Trackio as a New Backend for Experiment Visualization

[ParagEkbote]

Description

Trackio is a lightweight python library with a similar API to Wandb. It features an offline first design and the logging runs can be visualized as an HF Space or dataset on the HF Hub. I propose to add this backend as a callback to the the exisiting configurations similar to MLFlow, Comet, etc. I can help to integrate this backend to Ray, having done the same with optuna/optuna-integration#259. WDYT?

Use case

By adding trackio as a backend for experiment visualization, the users of Ray gain a local-first logging library, with the logging runs can be saved offline or on the HF Hub. This integration is particularly valuable for users already working within the Hugging Face ecosystem, as it bridges local experimentation with community‑driven sharing and collaboration.