Skip to content

hawkrobe/continual-adaptation

main
Switch branches/tags

Name already in use

A tag already exists with the provided branch name. Many Git commands accept both tag and branch names, so creating this branch may cause unexpected behavior. Are you sure you want to create this branch?
Code

Latest commit

 

Git stats

Files

Permalink
Failed to load latest commit information.
Type
Name
Latest commit message
Commit time
 
 
 
 
 
 
 
 
 
 

conventions

  • implementation of continual learning for repeated reference is in /models/
  • interactive, web-based reference game w/ model is in /behavioral_experiments/
  • experiments on model behavior are in /listener_experiments/ and /speaker_experiments/
  • csv output from both kinds of experiments goes in /data/
  • scripts for analyzing/visualizing data from experiments are in /analysis/
  • paper is in /writing/

To reproduce experiments

  1. install dependencies and the COCO API:
conda install pytorch torchvision -c pytorch
conda install Cython nltk
pip install -U symspellpy
git clone https://github.com/pdollar/coco.git
cd coco/PythonAPI/
make
python setup.py build
python setup.py install
cd ../../
  1. clone repo and download coco annotations and images for experiments
git clone https://github.com/hawkrobe/continual-adaptation.git
sh download-coco.sh
  1. Extract pretrained model weights and vocabulary files to ./data/preprocess/

For more information on pretrained model, see tutorial here

Dependencies

Model code depends on PyTorch >=1.2.0

About

Code and data for CoNLL paper "Continual Adaptation for Efficient Machine Communication"

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published