blobfish

The Blobfish Clickbait Detector by Davide Fonzi, Nicolò Pratelli e Lavinia Salicchi

Download and usage

1. Download training data and extract in the datasets /datasets/ subfolders:

   • /1/: clickbait16-train-170331.zip
   • /2/: clickbait17-train-170630.zip
   • /3/: clickbait17-unlabeled-170429.zip

2. git clone https://github.com/tira-io/blobfish

3. cd blobfish

4. install all dependencies for python 3.x

   pip3 install -r requirements.txt

5. for training

   python3 main.py

6. if you want to test our pretrained models

   python3 test.py -i test_data -o output -m model

   test_data: directory where is located test dataset

   output: path of the output text

   model: name of the choosen model (WordEmbNet, LingNet, FullNetPost, FullNet)

   if you want to test your own model, you must change path of the model into test.py

   dir = "longTraining/models/"
   if type_model == "FullNet":
    	model_path = dir + "path/to/model/model_name.hdf5"
   if type_model == "FullNetPost":
   	model_path = dir + "path/to/model/model_name.hdf5"
   if type_model == "LingNet":
   	model_path = dir + "path/to/model/model_name.hdf5"
   if type_model == "WordEmbNet":
   	model_path = dir + "path/to/model/model_name.hdf5"

7. for evaluating

   python3 eval.py test_data/truth.jsonl ./output/results.jsonl output.prototext