84d517d540204dcaa20237dc05bdd4a63a958923,reagent/training/td3_trainer.py,TD3Trainer,train,#TD3Trainer#Any#,102

Before Change

            next_actor = (next_actor + noise.clamp(*self.noise_clip_range)).clamp(
                *CONTINUOUS_TRAINING_ACTION_RANGE
            )
            next_state_actor = rlt.PreprocessedStateAction.from_tensors(
                next_state, next_actor
            )
            next_q_value = self.q1_network_target(next_state_actor)

            if self.q2_network is not None:
                next_q_value = torch.min(

After Change

            next_actor = (next_actor + noise.clamp(*self.noise_clip_range)).clamp(
                *CONTINUOUS_TRAINING_ACTION_RANGE
            )
            next_state_actor = (next_state, rlt.FeatureData(next_actor))
            next_q_value = self.q1_network_target(*next_state_actor)

            if self.q2_network is not None:
                next_q_value = torch.min(

Instances