pub struct ModelCompareResult {Show 14 fields
pub id: Option<i64>,
pub model_id: i64,
pub run_id: Option<i64>,
pub prompt_text: String,
pub system_prompt: Option<String>,
pub response_text: String,
pub was_truncated: bool,
pub prompt_tokens: Option<i64>,
pub completion_tokens: Option<i64>,
pub prompt_ms: Option<f64>,
pub generation_ms: Option<f64>,
pub prompt_tps: Option<f64>,
pub generation_tps: Option<f64>,
pub created_at: DateTime<Utc>,
}Expand description
Result of running a single model through a compare (inference) benchmark.
All timing fields are Option<f64> — llama-server may omit the timings
object. Missing values are stored as NULL in the database and surfaced as
None in the API; they never cause a panic.
Fields§
§id: Option<i64>Database ID of this result row (set after persistence).
model_id: i64Foreign key → models.id.
run_id: Option<i64>Foreign key → benchmark_runs.id (nullable; SET NULL on run delete).
prompt_text: StringPrompt text sent to this model.
system_prompt: Option<String>Optional system prompt.
response_text: StringFull response text accumulated from the stream.
was_truncated: booltrue if the response was cut short (finish_reason == "length").
prompt_tokens: Option<i64>Number of prompt tokens reported by the model.
completion_tokens: Option<i64>Number of completion tokens reported by the model.
prompt_ms: Option<f64>Time spent processing the prompt (milliseconds).
generation_ms: Option<f64>Time spent generating the response (milliseconds).
prompt_tps: Option<f64>Prompt-processing throughput (tokens/sec). None if timings absent.
generation_tps: Option<f64>Token-generation throughput (tokens/sec). None if timings absent.
created_at: DateTime<Utc>UTC timestamp of this result.
Trait Implementations§
Source§impl Clone for ModelCompareResult
impl Clone for ModelCompareResult
Source§fn clone(&self) -> ModelCompareResult
fn clone(&self) -> ModelCompareResult
1.0.0 (const: unstable) · Source§fn clone_from(&mut self, source: &Self)
fn clone_from(&mut self, source: &Self)
source. Read more