This model is ANCE-Tele trained on MS MARCO. The training details and evaluation results are as follows:
| Model | Pretrain Model | Train w/ Marco Title | Marco Dev MRR@10 | BEIR Avg NDCG@10 | 
|---|---|---|---|---|
| ANCE-Tele | cocodr-base | w/o | 37.3 | 44.2 | 
| BERI Dataset | NDCG@10 | 
|---|---|
| TREC-COVID | 77.4 | 
| NFCorpus | 34.4 | 
| FiQA | 29.0 | 
| ArguAna | 45.6 | 
| Touché-2020 | 22.3 | 
| Quora | 85.8 | 
| SCIDOCS | 14.6 | 
| SciFact | 71.0 | 
| NQ | 50.5 | 
| HotpotQA | 58.8 | 
| Signal-1M | 27.2 | 
| TREC-NEWS | 34.7 | 
| DBPedia-entity | 36.2 | 
| Fever | 71.4 | 
| Climate-Fever | 17.9 | 
| BioASQ | 42.1 | 
| Robust04 | 41.4 | 
| CQADupStack | 34.9 | 
The implementation is the same as our EMNLP 2022 paper "Reduce Catastrophic Forgetting of Dense Retrieval Training with Teleportation Negatives". The associated GitHub repository is available at https://github.com/OpenMatch/ANCE-Tele.
@inproceedings{sun2022ancetele,
  title={Reduce Catastrophic Forgetting of Dense Retrieval Training with Teleportation Negatives},
  author={Si, Sun and Chenyan, Xiong and Yue, Yu and Arnold, Overwijk and Zhiyuan, Liu and Jie, Bao},
  booktitle={Proceedings of EMNLP 2022},
  year={2022}
}
- Downloads last month
 - 1