Systems and efficient execution · MLD-0238
CPU offloading
Moving selected model or training states to host memory to reduce accelerator memory pressure.
Original reference page · detailed expansion pending. Original definitions, examples, notes and citations are preserved below. New examples are illustrative, not benchmark results.
Original example
Offload optimizer states during large-model training.
Original caveat
Transfers can become a major performance bottleneck.
Context and aliases
Level: Advanced. Stages: Training · Inference · Deployment.
Tags: DL · LLM · SLM · Systems.
Aliases: None recorded.
Related concepts
Original references
- PyTorch Fully Sharded Data Parallel ↗ · Consulted
- PyTorch documentation ↗ · Reference link
- vLLM documentation ↗ · Consulted
- Original reference URL ↗