Skip to content

Python-based and OpenAI spec conforming reinforcement learning environment that allows agents to select AABBs of words in images using a distinct set of actions

Notifications You must be signed in to change notification settings

MariusDanner/text-localization-environment

 
 

Repository files navigation

text-localization-environment

Build Status

The environment we will use to train our text localization agent.

For now, we will stick to the structure of the environment as laid out in the base paper. We conform to the OpenAi-Interface.

Installation

To install the package, clone this repository, open a terminal in the top folder of the package and call:

pip install -e .

To uninstall call

pip uninstall text_localization_environment

To use the environment, call

from text_localization_environment import TextLocEnv

To download the ResNet-152 caffemodel (it isn't downloaded automatically) see link and save it where necessary (an error will tell you where if you try to create a TextLocEnv).

About

Python-based and OpenAI spec conforming reinforcement learning environment that allows agents to select AABBs of words in images using a distinct set of actions

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 100.0%