Question

我有：

from transformers import XLNetTokenizer, XLNetForQuestionAnswering
import torch

tokenizer =  XLNetTokenizer.from_pretrained('xlnet-base-cased')
model = XLNetForQuestionAnswering.from_pretrained('xlnet-base-cased')

input_ids = torch.tensor(tokenizer.encode("What is my name?", add_special_tokens=True)).unsqueeze(0)  # Batch size 1
start_positions = torch.tensor([1])
end_positions = torch.tensor([3])
outputs = model(input_ids, start_positions=start_positions, end_positions=end_positions)
loss = outputs[0]

print(outputs)
print(loss)

根据文档。这可以提供一些帮助：

(tensor(2.3008, grad_fn=<DivBackward0>),)
tensor(2.3008, grad_fn=<DivBackward0>)

但是，如果可能，我想要一个实际的答案？

Answer 1

感谢Joe Davison提供了答案on Twitter：

from transformers import pipeline

qa = pipeline('question-answering')
response = qa(context='I like to eat apples, but hate bananas.',
              question='What do I like?')

print(response)

给出以下答复：

{'score': 0.282511100858045, 'start': 31, 'end': 38, 'answer': 'bananas.'}

不太正确，但至少得分很低。

如何实现拥抱面的基本回答？

1 个答案: