Sequence Parallelism: Long Sequence Training from System Perspective
Paper
•
2105.13120
•
Published
•
5
Long context LLM
-
-
Ring Attention with Blockwise Transformers for Near-Infinite Context
Paper • 2310.01889 • Published • 8 -
Striped Attention: Faster Ring Attention for Causal Transformers
Paper • 2311.09431 • Published • 4 -
DeepSpeed Ulysses: System Optimizations for Enabling Training of Extreme Long Sequence Transformer Models
Paper • 2309.14509 • Published • 16 -
LightSeq: Sequence Level Parallelism for Distributed Training of Long Context Transformers
Paper • 2310.03294 • Published • 2 -
BurstAttention: An Efficient Distributed Attention Framework for Extremely Long Sequences
Paper • 2403.09347 • Published • 20 -
Beyond the Limits: A Survey of Techniques to Extend the Context Length in Large Language Models
Paper • 2402.02244 • Published • 1 -
Infinite-LLM: Efficient LLM Service for Long Context with DistAttention and Distributed KVCache
Paper • 2401.02669 • Published • 10 -
Advancing Transformer Architecture in Long-Context Large Language Models: A Comprehensive Survey
Paper • 2311.12351 • Published • 3 -
LLM Maybe LongLM: Self-Extend LLM Context Window Without Tuning
Paper • 2401.01325 • Published • 24 -
LongRoPE: Extending LLM Context Window Beyond 2 Million Tokens
Paper • 2402.13753 • Published • 104 -
Leave No Context Behind: Efficient Infinite Context Transformers with Infini-attention
Paper • 2404.07143 • Published • 92 -
Megalodon: Efficient LLM Pretraining and Inference with Unlimited Context Length
Paper • 2404.08801 • Published • 62 -
Longformer: The Long-Document Transformer
Paper • 2004.05150 • Published • 3 -
Generating Long Sequences with Sparse Transformers
Paper • 1904.10509 • Published • 1 -
A Unified Sequence Parallelism Approach for Long Context Generative AI
Paper • 2405.07719 • Published • 1