Module lora 

Wire shapes for GET /lora-adapters, POST /lora-adapters and the per-request lora field, llama.cpp’s spellings.

The list GET returns is server_task_result_get_lora::to_json (tools/server/server-task.cpp:1608-1626) minus the aLoRA fields, which frink refuses to load. POST takes the same [{id, scale}] array parse_lora_request reads (server-common.cpp:131-142) and answers {"success": true} (server-task.cpp:1632-1634). The per-request lora field on a completion is that array again, and it means what construct_lora_list makes it mean (server-context.cpp:1721-1732): every adapter named gets its scale, every adapter NOT named gets 0 for that request.

Structs

LoraAdapterInfo

One loaded adapter, as GET /lora-adapters lists it.

LoraApplyResponse

The answer to POST /lora-adapters.

LoraScaleRequest

One entry of the [{id, scale}] array POST /lora-adapters and a request’s lora field carry.