Downloads · 30 days
0
wangsky/AR2-sieved-MS
AR2-sieved-MS is a machine learning model from wangsky. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as apache-2.0.
This is one of the trained models of the AAAI2024 paper: Mitigating the Impact of False Negatives in Dense Retrieval with Contrastive Confidence Regularization;
Downloads · 30 days
0
Access
Public
Updated Jan 8, 2024
Repo size
2.6 GB
Likes
0
Public
Click a slice to open those files.
Other2.6 GB · 100%
From the Hugging Face model README
This is one of the trained models of the AAAI2024 paper: Mitigating the Impact of False Negatives in Dense Retrieval with Contrastive Confidence Regularization;
You can directly use this model to produce the same results in our main experiments. You may refer to our github for how to use this model.
Evaluation on MS development set:
| MRR@5 | R@50 | R@1k | |
|---|---|---|---|
| AR2 | 39.5 | 87.8 | 98.6 |
| AR2+sieve (this model) | 39.8 | 88.3 | 98.6 |