Good work! I'm interested in how it performs on long-horizon memory embedding benchmark.

#3
by Yuki131 - opened
Memtensor Research Group org

Thank you for your interest. The MemReranker model is a reranking model, whereas the dataset you linked is a benchmark for evaluating embedding models. We are currently training an embedding model, which we plan to evaluate on this benchmark prior to its release. Please stay tuned for updates.

Sign up or log in to comment