# Lilt v0.1.0

Stage: **research baseline**. Not published to npm or Maven.

Authored multilingual sentences; last two sentences per language held out. Eight supported languages only.

## Artifact

- Raw bytes: 150048
- SHA-256: `6ed25e6ad7de36bcdce12b99c2a11bffd43a58e891c7f6595e9eaf6cf1d770d4`
- Training examples: 64
- Internal held-out smoke test: 15/16
- Classes: English, Spanish, French, German, Italian, Portuguese, Hindi, Japanese

The internal test is not an independent real-world benchmark. Scores are uncalibrated. Artifact size excludes the runtime and memory overhead.

## Node / browser

From the repository root: `bun install && bun run sdk:build`.

```ts
import { createModel } from '@embermote/lilt';
const model = createModel();
```

Text models accept a string. Audio models accept `audioFeatures(pcm, 8000)`; Glyph accepts `shapeFeatures(points)`, both exported by `@embermote/core`.

## Kotlin

```kotlin
import dev.embermote.lilt.Lilt
val model = Lilt.load()
```

JDK 17+, root Gradle build. Model resources are bundled in each model module. Use `Features.audio` or `Features.shape` for numeric inputs. Android device testing is pending.

## Rebuild

`bun run train` uses the original recipe in `training/train.py`. No external datasets or third-party weights are downloaded.