utils_nlp/models/transformers/named_entity_recognition.py [397:432]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
            device=device,
            num_gpus=num_gpus,
            max_steps=max_steps,
            gradient_accumulation_steps=gradient_accumulation_steps,
            optimizer=self.optimizer,
            scheduler=scheduler,
            fp16=fp16,
            amp=amp,
            local_rank=local_rank,
            verbose=verbose,
            seed=seed,
        )

    def predict(self, test_dataloader, num_gpus=None, gpu_ids=None, verbose=True):
        """
        Scores a dataset using a fine-tuned model and a given dataloader.

        Args:
            test_dataloader (DataLoader): DataLoader for scoring the data.
            num_gpus (int, optional): The number of GPUs to use.
                If None, all available GPUs will be used. If set to 0 or GPUs are
                not available, CPU device will be used.
                Defaults to None.
            gpu_ids (list): List of GPU IDs to be used.
                If set to None, the first num_gpus GPUs will be used.
                Defaults to None.
            verbose (bool, optional): Whether to print out the training log.
                Defaults to True.

        Returns
            1darray: numpy array of predicted label indices.
        """

        preds = list(
            super().predict(
                eval_dataloader=test_dataloader,
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -



utils_nlp/models/transformers/sequence_classification.py [323:358]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
            device=device,
            num_gpus=num_gpus,
            max_steps=max_steps,
            gradient_accumulation_steps=gradient_accumulation_steps,
            optimizer=self.optimizer,
            scheduler=scheduler,
            fp16=fp16,
            amp=amp,
            local_rank=local_rank,
            verbose=verbose,
            seed=seed,
        )

    def predict(self, test_dataloader, num_gpus=None, gpu_ids=None, verbose=True):
        """
        Scores a dataset using a fine-tuned model and a given dataloader.

        Args:
            test_dataloader (DataLoader): DataLoader for scoring the data.
            num_gpus (int, optional): The number of GPUs to use.
                If None, all available GPUs will be used. If set to 0 or GPUs are
                not available, CPU device will be used.
                Defaults to None.
            gpu_ids (list): List of GPU IDs to be used.
                If set to None, the first num_gpus GPUs will be used.
                Defaults to None.
            verbose (bool, optional): Whether to print out the training log.
                Defaults to True.

        Returns
            1darray: numpy array of predicted label indices.
        """

        preds = list(
            super().predict(
                eval_dataloader=test_dataloader,
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -