Tarek

Noticed the same problem for the `follow target` option with Brax v2. Followed the example shown [here](https://github.com/google/brax/blob/main/notebooks/training.ipynb). My pseudo-code looks like this: ```python env, pipeline_states = train(...) html_string = html.render(...

The last commit makes hyperparameter optimisation possible with the last version of optuna (v3.6.1). Related to #45. There might be uncaught errors left.

There seem to be an issue with the `if` statements on line [74](https://github.com/CarperAI/trlx/blob/main/trlx/trainer/accelerate_ppo_trainer.py#L74), [398](https://github.com/CarperAI/trlx/blob/main/trlx/trainer/accelerate_ppo_trainer.py#L398) and [424](https://github.com/CarperAI/trlx/blob/main/trlx/trainer/accelerate_ppo_trainer.py#L424) in the [trlx/trainer/accelerate_ppo_trainer.py](https://github.com/CarperAI/trlx/blob/main/trlx/trainer/accelerate_ppo_trainer.py) file. The check for `self.model.peft_type` should be made with `hasattr` like...