ML-QuantSubscribe

Machine learningOther

BAM! Just Like That: Simple and Efficient Parameter Upcycling for Mixture of Experts

The paper introduces BAM, a new method for training Mixture of Experts models, which optimizes the use of specialized dense models and improves performance in perplexity and downstream tasks.

Featured in No. 62 on 21 Aug 2024 · 6 days after release · 18 citations today · published in Neural Information Processing Systems

Released
15 Aug 2024
First featured
No. 62 · 21 Aug 2024
Citations (Semantic Scholar)
18
Influential citations
1
Published in
Neural Information Processing Systems
Shares when featured
47
Identifier
arXiv:2408.08274

Citations and venue from Semantic Scholar (ODC-BY), refreshed weekly. Summary: Quant Letter (CC BY 4.0).

    Type to search. Try rough volatility, LLM agents or FinGPT.

    ↑↓ move↵ openesc closeFull search page