Struct ModelTuning 

pub struct ModelTuning {
    pub model: String,
    pub num_ctx: Option<u32>,
    pub real_context_discovery: Option<bool>,
    pub mid_loop_trim_threshold: Option<usize>,
    pub mid_loop_trim_tokens: Option<usize>,
    pub max_tool_rounds: Option<usize>,
    pub workflow_grace_rounds: Option<usize>,
}

Inference-parameter overrides for a specific model name.

Matched against the active model by exact string equality. Add entries under [[model_tuning]] in ~/.newt/config.toml to pin parameters for models whose defaults cause problems (e.g. context overflow).

Human-authored entries are never touched by the auto-tuner. Auto-tuned entries are appended (not modified) when the harness gains high confidence in its empirical measurements.

Fields

model: String

Model name as it appears in Ollama (e.g. "nemotron3:33b").

num_ctx: Option<u32>

Ollama options.num_ctx — hard cap on KV-cache allocation. Overrides both the global [tui].num_ctx and the empirically derived safe_context from model-capabilities.json.

real_context_discovery: Option<bool>

Per-model override of [tui].real_context_discovery. Some(true) opts this model into empirical conserve-and-ratchet discovery; Some(false) forces trust-the-declared-window even when the global default is empirical; None inherits the global setting.

mid_loop_trim_threshold: Option<usize>

Per-model mid_loop_trim_threshold override.

mid_loop_trim_tokens: Option<usize>

Per-model mid_loop_trim_tokens override (estimated-token trim trigger).

max_tool_rounds: Option<usize>

Per-model max_tool_rounds override.

workflow_grace_rounds: Option<usize>

Per-model [tui].workflow_grace_rounds override.

Trait Implementations

impl Clone for ModelTuning

fn clone(&self) -> ModelTuning

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for ModelTuning

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for ModelTuning

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

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Eq for ModelTuning

impl PartialEq for ModelTuning

fn eq(&self, other: &ModelTuning) -> bool

Tests for self and other values to be equal, and is used by ==.

fn ne(&self, other: &Rhs) -> bool

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl Serialize for ModelTuning

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

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl StructuralPartialEq for ModelTuning