Options reference

Every option that any ext-infer method accepts, in one table per method. For conceptual context on individual options, follow the links in the rightmost column.

`Model::load($path, $options)`

The second argument is an associative array. Keys are kept as snake-case strings (like PHP ini settings) because load-time tuning is rare and the array form composes well with config arrays loaded from disk.

Key	Type	Default	See
`n_gpu_layers`	`int`	`0`	Performance tuning
`use_mmap`	`bool`	`true`	Performance tuning
`use_mlock`	`bool`	`false`	Performance tuning
`embedding`	`bool`	`false`	Embeddings
`pooling`	`string`	`'unspecified'`	Embeddings

Validation rules

Unknown keys are not rejected — they’re silently ignored. This is deliberate (forward-compatibility for callers loading config from files), but it means typos will be silent. If you suspect a typo, verify with var_dump against the same string before reporting a bug.
Type mismatches are rejected, with a clear message: invalid option n_gpu_layers: expected integer.
Negative integers and out-of-range values for n_gpu_layers are rejected: invalid option n_gpu_layers: must be non-negative.
pooling accepts only the six strings listed in Embeddings → Pooling.

`Model::chat($prompt, ...)`

Named arguments — no array. PHP 8.0+ named-arguments syntax echoes the ident verbatim, so you write maxTokens: 256 (camelCase, per PSR-12).

Argument	Type	Default	See
`$prompt`	`\Displace\Infer\Prompt`	required	Prompts
`maxTokens`	`int`	`128`	Chat completions
`nCtx`	`int`	`2048`	Chat completions
`temperature`	`float`	`0.0`	Chat completions
`seed`	`int`	`1234`	Chat completions

Behavior

temperature = 0.0 is greedy (deterministic). > 0.0 samples, controlled by seed.
seed is only consulted when temperature > 0.
maxTokens caps generation. Hitting it sets Response::finishReason() to 'length'.
nCtx is the context window for this call. If the rendered prompt exceeds it, InferenceException is raised before generation starts.

`Model::raw($prompt, ...)`

Same named-argument shape as chat() plus addBos.

Argument	Type	Default	See
`$prompt`	`string`	required	Raw completions
`maxTokens`	`int`	`128`	Chat completions
`nCtx`	`int`	`2048`	Chat completions
`temperature`	`float`	`0.0`	Chat completions
`seed`	`int`	`1234`	Chat completions
`addBos`	`bool`	`true`	Raw completions → addBos

`Model::embed($text)`

Just the text. Pooling and embedding-mode are configured at load time (see Model::load above).

Argument	Type	Default	See
`$text`	`string`	required	Embeddings

`Embedding` math

Embedding is read-only; the math methods return new instances rather than mutating.

Method	Returns
`vector()`	`list<float>`
`dimensions()`	`int`
`norm()`	`float`
`normalize()`	new `Embedding`
`cosineSimilarity(Embedding $other)`	`float` (in `[-1, 1]`)

cosineSimilarity throws InferenceException on a dimension mismatch — see Embeddings → vector math.

`Prompt`

Static factories + immutable with* builders.

Method	Returns
`Prompt::system($content)`	new `Prompt`
`Prompt::user($content)`	new `Prompt`
`withSystem($content)`	new `Prompt`
`withUser($content)`	new `Prompt`
`withAssistant($content)`	new `Prompt`
`messages()`	`list<Message>`
`lastRole()`	`?string`
`count()`	`int`
`isEmpty()`	`bool`

See Prompts for the immutability semantics.

`Response`

Read-only. Six getters.

Method	Returns
`text()`	`string`
`reasoning()`	`?string`
`answer()`	`string`
`hasReasoning()`	`bool`
`finishReason()`	`string` — `'eos'`/`'length'`/`'stop'`
`tokensGenerated()`	`int`

See Chat completions → Inspecting a Response.

Environment

Not strictly an option, but bears mentioning here:

Variable	Effect
`EXT_INFER_LOG=1`	Restore llama.cpp’s verbose stderr logging (silenced by default).

See Environment variables.

Keyboard shortcuts

ext-infer