| license: mit | |
| This model has been pretrained on BookCorpus and English Wikipedia following the approach described in the paper **Condenser: a Pre-training Architecture for Dense Retrieval**. The model can be used to reproduce the experimental results within the GitHub repository https://github.com/OpenMatch/COCO-DR. | |
| This model is trained with BERT-large as the backbone with 335M hyperparameters. |