oBERT-12-upstream-pruned-unstructured-97-finetuned-qqp-v2

This model is obtained with The Optimal BERT Surgeon: Scalable and Accurate Second-Order Pruning for Large Language Models.

It corresponds to the model presented in the Table 2 - oBERT - QQP 97% (in the upcoming updated version of the paper).

Pruning method: oBERT upstream unstructured + sparse-transfer to downstream
Paper: https://arxiv.org/abs/2203.07259
Dataset: QQP
Sparsity: 97%
Number of layers: 12

The dev-set performance reported in the paper is averaged over four seeds, and we release the best model (marked with (*)):

| oBERT 97%    | acc   | F1    |
| ------------ | ----- | ----- |
| seed=42   (*)| 90.42 | 87.09 |
| seed=3407    | 90.31 | 86.87 |
| seed=123     | 90.20 | 86.76 |
| seed=12345   | 90.39 | 87.16 |
| ------------ | ----- | ----- |
| mean         | 90.33 | 86.97 |
| stdev        | 0.098 | 0.186 |

Code: coming soon

BibTeX entry and citation info

@article{kurtic2022optimal,
  title={The Optimal BERT Surgeon: Scalable and Accurate Second-Order Pruning for Large Language Models},
  author={Kurtic, Eldar and Campos, Daniel and Nguyen, Tuan and Frantar, Elias and Kurtz, Mark and Fineran, Benjamin and Goin, Michael and Alistarh, Dan},
  journal={arXiv preprint arXiv:2203.07259},
  year={2022}
}