/multimodal_siamese_network

Primary LanguageCommon Lisp

Multimodal-Siamese-Network

This repository contains an implementation of Multimodal Object Identification using Siamese Network created in my graduation project.

Getting Started

We assume that you have access to a GPU with CUDA >= 11.0 support.

Installation

  1. Clone the repository
git clone https://github.com/makolon/Multimodal_Siamese_Network.git
  1. Build docker image

For robot control

cd ./robo_trainer/docker/
./build.sh

For object identification

cd ./docker
./build.sh
  1. Run docker container

For robot control

./robo_trainer/docker/
.run.sh

For object identification

cd ./docker
./run.sh

Usage

For robot control

cd Multimoda_Siamese_Network/robo_trainer/catkin_ws
source devel/setup.bash
roslaunch rs007n/rs007n_default.launch

For object identification

cd Multimoda_Siamese_Network/siamese_network/
python3 train.py <params>