Nodes/ComfyUI-cc_llama_vision/CC Llama Server Unload (free VRAM)
ComfyUI Node

CC Llama Server Unload (free VRAM)

A ComfyUI node in llama.cpp with 4 inputs and 1 output.

By cicalooo·Created 2 months ago·Updated about a month ago· 1
CC Llama Server Unload (free VRAM)
    • trigger
    trigger
    targetall_servers
    also_free_comfyui_vramtrue
    port8080
    Categoryllama.cpp

    Inputs (4)

    NameTypeDefaultDescription
    triggerSTRINGUpstream string output to chain from (e.g. the caption output) — guarantees this node runs before whatever consumes it, since ComfyUI executes in dependency order.
    targetCOMBOall_serversWhich server(s) to kill: every currently tracked llama-server process, or only the one on the specified port.
    also_free_comfyui_vramBOOLEANtrueAlso ask ComfyUI to unload its own loaded models and empty its VRAM cache after killing llama-server.
    portoptINT80801024–65535Port of the specific llama-server to kill, used only when 'target' is 'specific_port'.

    Outputs (1)

    NameTypeDescription
    triggerSTRING