hsthe29 / qazalo

Implementation of project available on Question Answering in Zalo AI Challenge

Bài toán Question Answering trong cuộc thi Zalo AI Challenge 2019

Thực tập kỹ thuật

@author: hsthe29

Overview

Given a question, related paragraphs from a Wikipedia article (in the shuffle order), the task is finding paragraph which answers the question of each test case.

F1 measure based on precision and recall is used to rank competing submissions.

For detailed, visit: https://challenge.zalo.ai/portal/question-answering

About Data?

Visit: https://challenge.zalo.ai/portal/question-answering

Approaches

I used pretrained bert (bert-base-multilingual-cased) for fine-tuning

Hyperparameter

num_train_epochs = 6.0

max_seq_length = 512

train_batch_size = 16

learning_rate = 2e-5

Run

Run the project using flags. For more details on the flags defined, read the main.py file or run this command: python main.py -h

Train

Please download the init checkpoint at here and put into the folder checkpoint.

Run command sh train.sh

Predict

If you want to only predict, please download the checkpoint of trained model at here

Run command: sh predict.sh.

Result:

Training logs are store at logs/.

For visualizing: $ tensorboard --logdir=logs/tensorboard

With validation set (data/train/val.csv), this model reach 83.6% of F1 score