How to make predictions from DistilBERT fine-tuned with PyTorch

Question

May I know how to continue from code of the link below, in order to reload the saved model and apply it on new data to make predictions? For a start, the model and token have been saved but not the JSON config file, so how do I generate it?

code

output_model_file = './models/pytorch_distilbert_news.bin'
output_vocab_file = './models/vocab_distilbert_news.bin'
model_to_save = model
torch.save(model_to_save, output_model_file)
tokenizer.save_vocabulary(output_vocab_file)

I have tried codes from various sources but was met with no success. Any help would be appreciated. :)