Libra bounds attention load-balancing to fixed-size sequence pools and combines variance-reduced sequence placement with tiled attention pooling, achieving up to 2.54x end-to-end throughput over Ulysses on 256K- and 1M-token Qwen3 training.
Title resolution pending
1 Pith paper cite this work, alongside 3 external citations. Polarity classification is still indexing.
1
Pith paper citing it
3
external citations · OpenAlex
fields
cs.DC 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Libra: Taming Attention Workload Skew in Long-Context LLM Training with Bounded Sequence Pool
Libra bounds attention load-balancing to fixed-size sequence pools and combines variance-reduced sequence placement with tiled attention pooling, achieving up to 2.54x end-to-end throughput over Ulysses on 256K- and 1M-token Qwen3 training.