1f720ee453871b2ab764f608926281716ef7bf81,tools/train_pl.py,LitModel,training_step,#LitModel#Any#Any#,88

Before Change

        logger_logs["scheduled_sampling_prob"] = self.model.ss_prob
        logger_logs["training_loss"] = loss

        output = {k:v if k == "loss" else v.data for k,v in model_out.items()}
        output["log"] = logger_logs

        return output

After Change

                                  data["gts"], torch.arange(0, len(data["gts"])), sc_flag, struc_flag)
        loss = model_out["loss"]

        data_time = self.trainer.profiler.recorded_durations["get_train_batch"][-1]
        data_time = torch.tensor(data_time)

        logger_logs = model_out.copy()
        if struc_flag or sc_flag:
            logger_logs["reward"] = model_out["reward"].mean()
            logger_logs["reward_var"] = model_out["reward"].var(1).mean()

        logger_logs["scheduled_sampling_prob"] = torch.tensor(
            self.model.ss_prob)
        logger_logs["training_loss"] = loss
        logger_logs["data_time"] = data_time

        output = {
            "loss": loss,
            "log": logger_logs,

Instances