Hongwu Peng

icon company University of Connecticut icon location CT. USA icon location Ph.D. candidate @ University of Connecticut

Results 2 issues of Hongwu Peng

Gradient caching vs Model dropout

3 comment

The GC-DPR has two steps 1. The first step did a full batch forward without gradient, to get the full batch contrastive learning loss and corresponding embedding gradient. 2. The...

MTEB STS Type's Dataset Generation

May I ask how you transform STS datasets into triplet? It seems like STS has a similarity score as ground truth, and how to include this information in the triplet...