giovanacampos/image-caption-generation

Image Caption Generation

Project based on the paper

---

https://www.cv-foundation.org/openaccess/content_cvpr_2015/papers/Vinyals_Show_and_Tell_2015_CVPR_paper.pdf

DevOps

---

Install

1. Install virtual environment on Linux

sudo apt install -y virtualenv

Setup

1. Setup virtual env. folder for python

virtualenv .venv

2. Activate environment sourcing

source .venv/bin/activate

3. Install requirements packages for code

pip install -r requirements.txt

Start

1. Activate environment sourcing

source .venv/bin/activate

Finish

1. Remove sourcing of virtual environment

deactive

Dataset

---

Pascal VOC 2008: 1000 test

• https://www.kaggle.com/sulaimannadeem/pascal-voc-2008

Flickr8k: 6000 train; 1000 valid; 1000 test

• https://www.kaggle.com/adityajn105/flickr8k

Flickr30k: 28000 train; 1000 valid; 1000 test

• https://www.kaggle.com/hsankesara/flickr-image-dataset

MSCOCO: 82783 train; 40504 valid; 40775 test

• https://cocodataset.org/#download

SBU: 1M train

• http://www.cs.virginia.edu/~vicente/sbucaptions/