Adacc: An Adaptive Framework Unifying Compression and Activation Recomputation for LLM Training

Documenti analoghi

• Optimizing Large Model Training through Overlapped Activation Recomputation
  di: Chen, Ping, et al.
  Pubblicazione: (2024)
• Frenzy: A Memory-Aware Serverless LLM Training System for Heterogeneous GPU Clusters
  di: Chang, Zihan, et al.
  Pubblicazione: (2024)
• HopGNN: Boosting Distributed GNN Training Efficiency via Feature-Centric Model Migration
  di: Chen, Weijian, et al.
  Pubblicazione: (2024)
• A Scene-aware Models Adaptation Scheme for Cross-scene Online Inference on Mobile Devices
  di: Li, Yunzhe, et al.
  Pubblicazione: (2024)
• MuxTune: Efficient Multi-Task LLM Fine-Tuning in Multi-Tenant Datacenters via Spatial-Temporal Backbone Multiplexing
  di: Xue, Chunyu, et al.
  Pubblicazione: (2026)