Exploiting Student Parallelism for Efficient GPU Inference of BERT-like Models in Online Services

Documenti analoghi

• BeamVQ: Aligning Space-Time Forecasting Model via Self-training on Physics-aware Metrics
  di: Wu, Hao, et al.
  Pubblicazione: (2024)
• Hydraulis: Balancing Large Transformer Model Training via Co-designing Parallel Strategies and Data Assignment
  di: Li, Haoyang, et al.
  Pubblicazione: (2024)
• Amoeba: Runtime Tensor Parallel Transformation for LLM Inference Services
  di: Chen, Haoyu, et al.
  Pubblicazione: (2025)
• BeamVQ: Beam Search with Vector Quantization to Mitigate Data Scarcity in Physical Spatiotemporal Forecasting
  di: Wang, Weiyan, et al.
  Pubblicazione: (2025)
• Exploiting Dependency and Parallelism: Real-Time Scheduling and Analysis for GPU Tasks
  di: Zhang, Yuanhai, et al.
  Pubblicazione: (2026)