BERT-QA

Building a machine reading comprehension system using pretrained model bert.

Quick Start

1. Prepare your training data and install the package in requirement.txt

2. Fine-tune BERT model

sh tarin.sh

3. Interaction

sh interaction.sh

Experiment

Input context format like below：

{ "sentence":"YOUR_SENTENCE。", "question":"YOUR_QUESTION"}

The experimental result of F1-measure：

Evaluation 100%|███████████████████████████████████| 495/495 [00:05<00:00, 91.41it/s]
Average f1 : 0.5596300989105781

Display Result

Model architectures

BERT (from Google) released with the paper BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding by Jacob Devlin, Ming-Wei Chang, Kenton Lee and Kristina Toutanova.

Dataset

In this experiments, we use the datasets from DRCKnowledgeTeam. (https://github.com/DRCKnowledgeTeam/DRCD)