The present invention discloses a reading comprehension model training method and a reading comprehension processing method, which relates to the fields of natural language processing and deep learning technology. A specific implementation method may include inputting the first training sample of the reference region into the reference reading comprehension model of the reference region and the target reading comprehension model of the target region, respectively, and first output data output by the reference reading comprehension model and second output data output by the target reading comprehension model The second output data is obtained, and a first training process is performed on the target reading comprehension model according to the difference between the first output data and the second output data, thereby causing the target reading comprehension model to directly apply the reference region knowledge already learned by the reference reading comprehension model. learning, and the target reading comprehension model avoids the process of learning the reference region knowledge from the first training sample of the reference region, so that the model training is more efficient, and the training result obtained by simply relying only on the first training sample of the reference region You can get better training results.
展开▼