📖 The AI Tool Bible
inference

Batching (Inference)

Grouping multiple requests into one forward pass to keep GPU utilisation high. Continuous batching lets requests join and leave mid-generation.

Related terms

Tools that implement Batching (Inference)