Skip to content

vllm.model_executor.model_loader.weight_cache

Modules:

  • daemon –

    Weight cache daemon for fast engine restarts.

  • ipc_loader –

    IPC model loader: maps post-quantized weights from a local weight cache

  • protocol –

    WeightCacheKey fingerprinting and socket protocol for the weight cache daemon.

  • utils –

    Helpers shared by the weight cache daemon, the IPC loader and the engine.