def get_prediction(sentence): sentence = Datasets.normalize_string(sentence) sentence = tokenizer.tokenize(sentence) sentence = tokenizer.convert_tokens_to_ids(sentence) sentence = [vocab['[CLS]']] + sentence + [vocab['[SEP]']] output = model(torch.tensor(sentence).unsqueeze(0)) output_softmax = softmax(output)[0] max_out = label_list[output_softmax.argmax()] argidx = output_softmax.argsort(descending=True) result = {label_list[i]: round(output_softmax[i].item(), 3) for i in range(len(label_list))} sorted_result = {label_list[i]: round(output_softmax[i].item(), 3) for i in argidx} return max_out, result, sorted_result