This repository contains the code and the model weights based on the validation loss for the evaluation task for contributing to the project "Evaluation of Language Models on Dark Web Text using BERT". The project is part of Google Summer of Code (GSoC).
The notebook is in the repository named gsoc-eval-test (1).ipynb
The weights of the best-performing models trained as part of this project can be found here.
To load the weights into the model, follow these steps:
import torch
# Replace "directory_weights" with the directory of the weights
weights = torch.load("directory_weights")
model.load_state_dict(weights["model"])