metadata

tags:
  - bert
  - oBERT
  - sparsity
  - pruning
  - compression
language: en
datasets: mnli

oBERT-12-downstream-pruned-unstructured-97-mnli

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 1 - 30 Epochs - oBERT - MNLI 97%.

Pruning method: oBERT downstream unstructured
Paper: https://arxiv.org/abs/2203.07259
Dataset: MNLI
Sparsity: 97%
Number of layers: 12

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

| oBERT 97%    | m-acc | mm-acc|
| ------------ | ----- | ----- |
| seed=42   (*)| 82.10 | 81.94 |
| seed=3407    | 81.81 | 82.27 |
| seed=54321   | 81.40 | 81.83 |
| ------------ | ----- | ----- |
| mean         | 81.77 | 82.01 |
| stdev        | 0.351 | 0.228 |

Code: https://github.com/neuralmagic/sparseml/tree/main/research/optimal_BERT_surgeon_oBERT

If you find the model useful, please consider citing our work.

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}
}