pi05-libero discrete_state_input = False

[chenkang455]

In the pi05-libero config, discrete_state_input is set to False, why is it designed in this way? In the libero, pi05-libero doesn't take the state as the input?

    TrainConfig(
        name="pi05_libero",
        model=pi0_config.Pi0Config(pi05=True, action_horizon=10, discrete_state_input=False),
        data=LeRobotLiberoDataConfig(
            repo_id="physical-intelligence/libero",
            base_config=DataConfig(prompt_from_task=True),
            extra_delta_transform=False,
        ),
        batch_size=256,
        lr_schedule=_optimizer.CosineDecaySchedule(
            warmup_steps=10_000,
            peak_lr=5e-5,
            decay_steps=1_000_000,
            decay_lr=5e-5,
        ),
        optimizer=_optimizer.AdamW(clip_gradient_norm=1.0),
        ema_decay=0.999,
        weight_loader=weight_loaders.CheckpointWeightLoader("gs://openpi-assets/checkpoints/pi05_base/params"),
        pytorch_weight_path="/path/to/your/pytorch_weight_path",
        num_train_steps=30_000,
    ),

If this is not set, this feature will be enabled by:

  def __post_init__(self):
      if self.max_token_len is None:
          object.__setattr__(self, "max_token_len", 200 if self.pi05 else 48)
      if self.discrete_state_input is None:
          object.__setattr__(self, "discrete_state_input", self.pi05)

In the pi05 model, the state info is encoded in the language encoding process, however, if this is set to false, pi05 will not process the state?