https://github.com/jazzsaxmafia/show_attend_and_tell.tensorflow

https://github.com/jazzsaxmafia/show_attend_and_tell.tensorflow

Last synced: 10 days ago
JSON representation

• Host: GitHub
• URL: https://github.com/jazzsaxmafia/show_attend_and_tell.tensorflow
• Owner: jazzsaxmafia
• License: bsd-2-clause
• Created: 2015-11-16T14:47:39.000Z (about 9 years ago)
• Default Branch: master
• Last Pushed: 2019-10-23T10:03:31.000Z (about 5 years ago)
• Last Synced: 2024-10-28T21:41:15.559Z (about 2 months ago)
• Language: Jupyter Notebook
• Size: 495 KB
• Stars: 507
• Watchers: 23
• Forks: 191
• Open Issues: 21
• Metadata Files:
  • Readme: Readme.md
  • License: LICENSE

Awesome Lists containing this project

• awesome-tensorflow - Neural Caption Generator with Attention - Implementation of ["Show, Attend and Tell"](http://arxiv.org/abs/1502.03044) (Models/Projects)
• awesome-tensorflow - Neural Caption Generator with Attention - Implementation of ["Show, Attend and Tell"](http://arxiv.org/abs/1502.03044) (Models/Projects)
• Awesome-TensorFlow-Chinese - Neural Caption Generator with Attention - Implementation of ["Show, Attend and Tell"](http://arxiv.org/abs/1502.03044) (模型项目 / 微信群)
• fucking-awesome-tensorflow - Neural Caption Generator with Attention - Implementation of ["Show, Attend and Tell"](http://arxiv.org/abs/1502.03044) (Models/Projects)

README

        
# Neural Caption Generator with Attention

* Tensorflow implementation of "Show, attend and Tell" http://arxiv.org/abs/1502.03044

* Borrowed most of the idea from the author's source code https://github.com/kelvinxu/arctic-captions

## Code

* make_flickr_dataset.py: Extracts conv5_3 layer activations of VGG Network for flickr30k images, and save them in 'data/feats.npy'

* model_tensorflow.py: Main codes

## Usage

* Download flickr30k Dataset.

* Extract VGG conv5_3 features using make_flickr_dataset.py

* Train: run train() in model_tensorflow.py

* Test: run test() in model_tensorflow.py

![alt tag](https://github.com/jazzsaxmafia/show_attend_and_tell.tensorflow/blob/master/attend.jpg)