Systems and efficient execution · MLD-0434
Float16
A 16-bit floating-point format with lower storage cost but a narrower numeric range than float32.
Original reference page · detailed expansion pending. Original definitions, examples, notes and citations are preserved below. New examples are illustrative, not benchmark results.
Original example
Reduce activation memory and accelerate supported operations.
Original caveat
Its limited range can cause overflow or underflow.
Context and aliases
Level: Advanced. Stages: Training · Inference · Deployment.
Tags: DL · LLM · SLM · Systems.
Aliases: FP16 · Half precision.
Related concepts
Concepts linking here
Original references
- PyTorch Fully Sharded Data Parallel ↗ · Consulted
- PyTorch documentation ↗ · Reference link
- vLLM documentation ↗ · Consulted
- Original reference URL ↗