Struct ModelPerfResult 

pub struct ModelPerfResult {
    pub id: Option<i64>,
    pub model_id: i64,
    pub run_id: Option<i64>,
    pub pp_tps: f64,
    pub tg_tps: f64,
    pub pp_tokens: i64,
    pub tg_tokens: i64,
    pub backend: Option<String>,
    pub ngl: Option<i64>,
    pub context_size: Option<i64>,
    pub repetitions: i64,
    pub created_at: DateTime<Utc>,
}

Result of running llama-bench on a single model.

Fields

id: Option<i64>

Database ID of this result row (set after persistence).

model_id: i64

Foreign key → models.id.

run_id: Option<i64>

Foreign key → benchmark_runs.id (nullable; SET NULL on run delete).

pp_tps: f64

Prompt-processing throughput (tokens/sec).

tg_tps: f64

Token-generation throughput (tokens/sec).

pp_tokens: i64

Number of prompt tokens used in the benchmark.

tg_tokens: i64

Number of generation tokens used in the benchmark.

backend: Option<String>

Backend reported by llama-bench (e.g. “Metal”, “CUDA”, “CPU”).

ngl: Option<i64>

Number of GPU layers offloaded.

context_size: Option<i64>

Context size used.

repetitions: i64

Number of repetitions averaged.

created_at: DateTime<Utc>

UTC timestamp of this result.

Trait Implementations

impl Clone for ModelPerfResult

fn clone(&self) -> ModelPerfResult

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for ModelPerfResult

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for ModelPerfResult

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Serialize for ModelPerfResult

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.