Skip to content

tapunict/hypermeme

 
 

Repository files navigation

hypermeme

======= inspired by NicsMeme®

"In the future memes will be able to generate themselves and propagate automatically"

see also the Dead Internet Theory

Meme competition with Nics


For a complete presentation of this project see hypermeme.ipynb

Project overview

Brief description

The aim of this project is to categorize memes into 4 categories or topics.

  • pol: politics
  • ent: entertainment
  • sport: sports
  • oth: other

The classification is based on their visual content (ocr and caption).

Project structure

Project structure

Setup

  1. Make sure you got docker and wget installed on your machine as they are required for this project

  2. Download Kafka

    cd ./kafka/setup
    wget https://downloads.apache.org/kafka/3.7.1/kafka_2.13-3.7.1.tgz 
  3. Build containers

    # if you want gpu acceleration
    docker compose -f gpu_compose.yaml build
    # otherwise (cpu)
    docker compose build
  4. Start pipeline (see quickstart)

  5. Import dashboard+data_view.ndjson file from /kibana directory into kibana

Quickstart

    # if you want gpu acceleration
    docker compose -f gpu_compose.yaml --profile pipeline up
    # otherwise (cpu)
    docker compose --profile pipeline up

Download the dataset

    docker compose --profile download_dataset up

Model training

    # Pretrained model is already included with this repo. If you want to rebuild it using your own data you can use this command
    docker compose --profile build_model up

Useful links

Container URL Description
kafka-UI http://localhost:8080 Open kafka UI
kibana http://localhost:5601 Kibana base URL

Dashboard demo

Dashboard

Technologies Used

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • HTML 71.6%
  • Python 14.7%
  • Jupyter Notebook 12.9%
  • Other 0.8%