cs8524-pictureteller

This project lets user upload images and return a related sentence.

Getting Started

Prerequisites

Java 8
Maven
Python 2.7
Numpy
PyTorch
pycocotools

Built With

• Maven - Dependency Management

Download Pretrained model and Vocabulary

If you do not want to train the model from scratch, you can use a pretrained model. You can download the pretrained model here and the vocabulary file here. You should extract pretrained_model.zip to ./image_caption/models/ and vocab.pkl to ./image_caption/data/ using unzip command.

How to run an example?

First of all, git clone --recurse-submodules -j8 our project and cd to the current project,

git clone --recurse-submodules -j8 https://github.com/yaodongyu/cs8524-pictureteller.git

cd cs8524-pictureteller

0. Download Pretrained model and Vocabulary

Download the pretrained model here and the vocabulary file here.

Extract pretrained_model.zip to ./image_caption/models/ and vocab.pkl to ./image_caption/data/ using unzip command.

1. Package this project

mvn package

2. Move the .jar file under cs8524-pictureteller/

mv target/pictureteller-0.0.1-SNAPSHOT.jar .

3. Deployment

java -jar pictureteller-0.0.1-SNAPSHOT.jar

4. Open a browser and have fun

1). To browse the landing page, type

localhost:8080

2). To initialize a new user, type

localhost:8080/user/new

3). To add an image to user #id, such as user 1, type

localhost:8080/user/1/image

4). To show an image of user #id, such as user 1, type

localhost:8080/user/1/show

5. How to use the h2 database?

1). Open h2 console

localhost:8080/h2-console

2). Enter user name and related information, where

spring.datasource.driver-class-name=org.h2.Driver
spring.datasource.url=jdbc:h2:file:c:/temp/rcp_h2
spring.datasource.username=sa
spring.datasource.password=