Logical Fallacy Detection

A repository containing the network to detect and classify logical fallacies in text.

Model evaluation

Binary classification

Multiclass classification

• Run docker
• Modify docker settings to:

{
  "builder": {
    "gc": {
      "defaultKeepStorage": "20GB",
      "enabled": true
    }
  },
  "experimental": false,
  "features": {
    "buildkit": false
  }
}

• Build and run docker

chmod +x deploy.sh
./deploy.sh

Repo structure

• network/models - Models weights.
• network/utils - Models definitions and utils for inference.
• network/training/data - Data used for training binary and multiclass classification models.
• network/training/baseline - Code for training baseline NLI models.
• network/training/prototex - Code for training CBR ProtoTEx model.