Training Loss Function - Githubissues

castorini / rank_llm

RankLLM is a Python toolkit for reproducible information retrieval research using rerankers, with a focus on listwise reranking.

http://rankllm.ai

Apache License 2.0

312 stars 39 forks source link

Closed melonking32 closed 8 months ago

melonking32 commented 10 months ago

What is the loss function you use during training, LM loss, or ranknet loss？

ronakice commented 8 months ago

We train our models with the standard teacher forcing objective.