This is a model checkpoint for "[Structured Pruning Learns Compact and Accurate Models](https://arxiv.org/pdf/2204.00408.pdf)". The model is pruned from `bert-base-uncased` to a 95% sparsity on dataset MNLI. Please go to [our repository](https://github.com/princeton-nlp/CoFiPruning) for more details on how to use the model for inference. Note that you would have to use the model class specified in our repository to load the model.