🎯 Trackio integration (#38814)

* First attempt

* fix

* fix

* Enhance TrackioCallback to log GPU memory usage and allocation

* Enhance Trackio integration in callbacks and training arguments documentation

* re order

* remove unused lines

* fix torch optional

docs/source/en/main_classes/callback.md

@@ -33,6 +33,7 @@ By default, `TrainingArguments.report_to` is set to `"all"`, so a [`Trainer`] wi
   it's the second one).
 - [`~integrations.TensorBoardCallback`] if tensorboard is accessible (either through PyTorch >= 1.4
   or tensorboardX).
+- [`~integrations.TrackioCallback`] if [trackio](https://github.com/gradio-app/trackio) is installed.
 - [`~integrations.WandbCallback`] if [wandb](https://www.wandb.com/) is installed.
 - [`~integrations.CometCallback`] if [comet_ml](https://www.comet.com/site/) is installed.
 - [`~integrations.MLflowCallback`] if [mlflow](https://www.mlflow.org/) is installed.
@@ -72,6 +73,9 @@ Here is the list of the available [`TrainerCallback`] in the library:
 
 [[autodoc]] integrations.TensorBoardCallback
 
+[[autodoc]] integrations.TrackioCallback
+    - setup
+
 [[autodoc]] integrations.WandbCallback
     - setup