DeFT: Flash Tree-attention with IO-Awareness for Efficient Tree-search-based LLM Inference.
Jinwei YaoKaiqi ChenKexun ZhangJiaxuan YouBinhang YuanZeke WangTao LinPublished in: CoRR (2024)
Keyphrases
- tree search
- search algorithm
- constraint propagation
- mathematical programming
- branch and bound
- tree search algorithm
- search tree
- alpha beta
- game tree search
- game tree
- tree structure
- depth first search
- iterative deepening
- path finding
- cost function
- special case
- search space
- database systems
- decision trees
- genetic algorithm