HSAP introduces a hierarchical framework and sequence-aware algorithm with JIT-optimized NCCL communication to enable correct causal attention computation on hybrid-context packed sequences without limiting parallelism.
arXiv preprint arXiv:2311.02382 , year=
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
HSAP: A Hierarchical Sequence-aware Parallelism for Hybrid-Context Generative Models
HSAP introduces a hierarchical framework and sequence-aware algorithm with JIT-optimized NCCL communication to enable correct causal attention computation on hybrid-context packed sequences without limiting parallelism.