In this example, you learn how to use the BERT QA model trained by GluonNLP (Apache MXNet) and PyTorch. You can provide the model with a question and a paragraph containing an answer. The model is then able to find the best answer from the answer paragraph. You can find the source code in BertQaInference.java.
Note that Apache MXNet BERT model has a limitation where the max size of the tokens including the question and the paragraph is 384.
Example:
Q: When did BBC Japan start broadcasting?
Answer paragraph:
BBC Japan was a general entertainment channel, which operated between December 2004 and April 2006.
It ceased operations after its Japanese distributor folded.
And it picked the right answer:
A: December 2004
Follow setup to configure your development environment.
cd examples
./gradlew run -Dmain=ai.djl.examples.inference.nlp.BertQaInference