https://github.com/astridnielsen-lab/imageslegend

https://github.com/astridnielsen-lab/imageslegend

Last synced: 17 days ago
JSON representation

• Host: GitHub
• URL: https://github.com/astridnielsen-lab/imageslegend
• Owner: AstridNielsen-lab
• License: mit
• Created: 2021-04-18T19:38:08.000Z (almost 4 years ago)
• Default Branch: master
• Last Pushed: 2021-04-18T19:38:18.000Z (almost 4 years ago)
• Last Synced: 2024-11-19T18:01:40.534Z (3 months ago)
• Language: Python
• Size: 267 KB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# medium-show-and-tell-caption-generator

Code to run inference on a Show And Tell Model.

This model is trained to generate captions given an image.

## Weblinks

A detailed tutorial for the code can be found here:

https://medium.freecodecamp.org/building-an-image-caption-generator-with-deep-learning-in-tensorflow-a142722e9b1f

## Getting started

1. Download Docker image

```bash

$ sudo docker pull colemurray/medium-show-and-tell-caption-generator

```

2. Download model graph and weights

```bash

$ sudo docker run -e PYTHONPATH=$PYTHONPATH:/opt/app -v $PWD:/opt/app \

-it colemurray/medium-show-and-tell-caption-generator \

python3 /opt/app/bin/download_model.py \

--model-dir /opt/app/etc

```

3. Download vocabulary

```bash

$ curl -o etc/word_counts.txt https://raw.githubusercontent

.com/ColeMurray/medium-show-and-tell-caption-generator/master/etc/word_counts.txt

```

4. Add testing images to ```imgs/``` folder

5. Generate captions

```bash

$ sudo docker run -v $PWD:/opt/app \

-e PYTHONPATH=$PYTHONPATH:/opt/app \

-it colemurray/medium-show-and-tell-caption-generator  \

python3 /opt/app/medium_show_and_tell_caption_generator/inference.py \

--model_path /opt/app/etc/show-and-tell.pb \

--vocab_file /opt/app/etc/word_counts.txt \

--input_files /opt/app/imgs/*

```

## Run with conda environment

These steps assume that model and vocabulary are already downloaded.

1. Build and activate conda environment

```bash

$ conda env create -f environment.yml

$ source activate show-and-tell

```

2. Run inference

```bash

$ python -m medium_show_and_tell_caption_generator.inference --model_path='etc/show-and-tell.pb' \

--vocab_file='etc/word_counts.txt' \

--input_files='imgs/*'

```