AttributeError: 'MMBTConfig' object has no attribute 'use_return_dict'

[Tdrinker]

Describe the bug I encountered AttributeError: 'MMBTConfig' object has no attribute 'use_return_dict' when using MultiModalClassificationModel.

# Create a ClassificationModel
args = {'reprocess_input_data': True, 
        'overwrite_output_dir': True,
        'fp16': False,
        'n_gpu': 2,
        'save_model_every_epoch': False,
        'evaluate_during_training': True,
        'num_train_epochs': max_train_epochs,
        'min_frequency': 5,
        'train_batch_size': 128,
        'best_model_dir': best_model_dir,
       }

model = MultiModalClassificationModel(model_type="bert", model_name="bert-base-uncased", label_list=[0, 1, 2, 3, 4, 5, 6, 7],  args=args)

model.train_model(nemo_data_splits['train'], eval_data=nemo_data_splits['val'], image_path='all_shooting_images_newest/')

and my nemo_data_splits looks like:

Additional context Potentially similar related issue: https://github.com/flairNLP/flair/issues/1841

[stale[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.