Struct anira::PlanQuery#

struct PlanQuery#

Collaboration diagram for anira::PlanQuery:

digraph {
    graph [bgcolor="#00000000"]
    node [shape=rectangle style=filled fillcolor="#FFFFFF" font=Helvetica padding=2]
    edge [color="#1414CE"]
    "1" [label="anira::PlanQuery" tooltip="anira::PlanQuery" fillcolor="#BFBFBF"]
    "2" [label="anira::PlanQuery::Backend" tooltip="anira::PlanQuery::Backend"]
    "1" -> "2" [dir=forward tooltip="usage"]
}

What PlanReport::find looks for, the anira_plan_query of anira_plan_report_find as an aggregate: every field names one property a plan must have, or matches every plan. report.find({.backend = {.engine = ANIRA_ENGINE_ONNXRUNTIME}}) finds the first plan on ONNX Runtime, report.find({.backend = {.provider = ANIRA_PROVIDER_COREML}}) the first on Core ML whatever its engine, report.find({.backend = {.engine = ANIRA_ENGINE_CUSTOM, .engine_id = "com.example.engine"}}) the first on that custom engine.

Public Members

uint32_t variant = ANIRA_VARIANT_ANY#

The variant of the inference stage; ANIRA_VARIANT_ANY matches every variant.

Backend backend = {}#
struct Backend#

The backend axes; NONE matches every engine or every provider. A custom engine is ANIRA_ENGINE_CUSTOM with its engine_id, a custom provider ANIRA_PROVIDER_CUSTOM with its provider_id (the pair rule: an id exactly beside its CUSTOM value).

Public Members

EngineKind engine = ANIRA_ENGINE_NONE#
Provider provider = ANIRA_PROVIDER_NONE#
const char *engine_id = nullptr#
const char *provider_id = nullptr#