KnapFormer claims knapsack-based token rebalancing gives <1% workload gap and 2x-3x faster diffusion-transformer training, but the attached full text is a different paper about physics simulations.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.DC 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
KnapFormer: An Online Load Balancer for Efficient Diffusion Transformers Training
KnapFormer claims knapsack-based token rebalancing gives <1% workload gap and 2x-3x faster diffusion-transformer training, but the attached full text is a different paper about physics simulations.