nthngdy's picture
Update README.md
bfb06ff
|
raw
history blame
873 Bytes
metadata
license: mit
datasets:
  - the_pile_openwebtext2
language:
  - en
pipeline_tag: text-generation

Model Sources

Model Architecture and Objective

This model is a Pythia-70m architecture trained on OpenWebText-2 using the Contrastive Weight Tying objective, and briefly fine-tuned for language generation on the same dataset.

Citation

BibTeX:

@misc{godey2023headless,
      title={Headless Language Models: Learning without Predicting with Contrastive Weight Tying}, 
      author={Nathan Godey and Éric de la Clergerie and Benoît Sagot},
      year={2023},
      eprint={2309.08351},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}

Contact

nathan.godey@inria.fr