SKaiNET aims to democratize "Edge AI / On-device AI" by bridging the gap between high-level application development and low-level hardware optimization. We believe AI should be portable, type-safe, and developer-friendly, enabling seamless intelligence in everything from mobile apps to IoT devices without sacrificing performance.
For architecture details see ARCHITECTURE.md.
Add the core dependencies (Gradle Kotlin DSL):
dependencies {
// Recommended: import the umbrella BOM and drop versions on the engine modules.
implementation(platform("sk.ainet:skainet-bom:0.23.0"))
implementation("sk.ainet.core:skainet-lang-core")
implementation("sk.ainet.core:skainet-backend-cpu")
}The BOM was first correctly published to Maven Central in 0.22.2 — earlier versions shipped at the wrong coordinates and could not be imported. Pin versions directly if you need an older release.
val model = nn {
input(28 * 28)
dense(out = 128)
relu()
dense(out = 10)
}val a = tensor(shape(2, 2)) { float(1f, 2f, 3f, 4f) }
val b = tensor(shape(2, 2)) { float(5f, 6f, 7f, 8f) }
val c = a matMul b
val d = c.relu()// Recommended: streaming reader — memory-efficient, supports quantized types
val source = JvmRandomAccessSource.open("model.gguf")
StreamingGGUFReader.open(source).use { reader ->
println("Tensors: ${reader.tensorCount}")
// Load specific tensor on demand (no whole-file loading)
val bytes = reader.loadTensor("token_embd.weight")
// Or get a TensorStorage descriptor with encoding/placement metadata
val storage = reader.loadTensorStorage("token_embd.weight")
}More examples: SKaiNET-examples | SKaiNET-notebook
SKaiNET is a modular ecosystem. While this repository contains the core engine, specialized high-level libraries are maintained in standalone repositories:
| Project | Description |
|---|---|
| SKaiNET-transformers | Pre-built transformer architectures and layers |
| SKaiNET-examples | Sample projects and integration demos |
| Goal | Start here |
|---|---|
| Examples and sample projects | SKaiNET-examples |
| Interactive notebooks | SKaiNET-notebook |
| LLM inference (Llama, Gemma, Qwen) | SKaiNET-transformers |
- Targets: JVM, macOS (Native), JS, WASM (Browser + WasmWasi)
- Single codebase shared across all platforms via Kotlin Multiplatform
- ComputeGraphExecutor: Optimized engine with fusion passes and trace-to-DAG bridging.
- SDPA & Gather: High-performance Scaled Dot-Product Attention and indexing operations.
- TurboQuant: Runtime KV-cache compression (~8x at 4-bit) for long-context LLM inference. Presets:
safe-lowbit,balanced,experimental-max. SeeTurboQuantUsagefor integration guide.
- ComputeGraph: Unified framework for defining agentic workflows and tool-calling loops.
- Java facade:
JavaAgentLoop(inskainet-lang-java)
- Sequential:
nn { input(); dense(); relu(); dense() } - DAG / Graph: arbitrary wiring with
dag { }for ResNet, YOLO-style architectures - Layers: Dense, Conv1d/2d/3d, MaxPool, AvgPool, BatchNorm, Dropout, LeakyReLU, ELU
- KAN (Kolmogorov–Arnold Networks) layer (experimental)
- Autograd engine with reverse-mode gradients, SGD and Adam/AdamW optimizers
- Built-in loaders: MNIST, Fashion-MNIST, CIFAR-10
- Formats: GGUF, ONNX, SafeTensors, JSON, Image (JPEG, PNG)
- Type-safe transform DSL: resize, crop, normalize, toTensor
SKaiNETentry point,TensorJavaOps, builder-pattern model definition- Maven BOM (
sk.ainet:skainet-bom) for one-line version management
- Export trained models to standalone, optimized C99 with static memory allocation
- Ready-to-use Arduino library output
- Lower Kotlin DSL to MLIR StableHLO dialect
- Optimization passes: constant folding, operation fusion, dead code elimination
- Valid IREE-compilable output with streaming API and public
HloGenerator
- Real-model GGUFs no longer OOM at network construction. The DSL pre-allocated zero-filled
FloatArray(shape.volume)for every Linear / Conv weight at module-creation time, even though downstream loaders overwrite those zeros immediately. For an Apertus-8B Q4_K_S GGUF (4.7 GB on disk) that was ~27 GB of FP32 zeros allocated and thrown away — OOMed at 12 GB heap. NewTensorDataFactory.placeholder(...)API; every eagerzeros(...)call site in the network builders routes through it. Lazy materialization fires only if a caller actually reads the tensor (which the load path never does). Verified end-to-end againstunsloth/Apertus-8B-Instruct-2509-GGUF: now loads in 12 GB heap. Same fix benefits Gemma / Llama / Qwen / Voxtral DSL paths transparently. (Issue #587, PR #588) - Kotlin/Native: GGUFs over ~2 GiB now load.
createRandomAccessSource(filePath)had no native actual; K/N consumers fell through to the legacy slurp-into-ByteArrayreader, which capped atInt.MAX_VALUEbytes (~2 GiB). Practical impact: macOS / Linux / iOS native couldn't open Q8 models above ~1B parameters or Q4 above ~3B. New POSIX-pread-backedPosixPreadRandomAccessSourcecoversmacosArm64,linuxX64,linuxArm64,iosArm64,iosSimulatorArm64. (Issue #589, PR #591)
- 0.22.2 —
sk.ainet:skainet-bomnow resolves from Maven Central (earlier versions shipped at the wrong coordinates). (Issue #584) - 0.22.1 —
StreamingShardedSafeTensorsReader.loadTensorStorageMappedfor zero-copy reads of multi-shard tensors above the 2 GB JVMByteArraylimit. (PR #582) - 0.22.0 — Native (FFM) CPU kernel provider: 4–6× faster Q4_K matmul, 1.5–1.8× FP32 SGEMM vs Panama Vector; auto-selected via
KernelRegistry.bestAvailable(). (PR #571)
See CHANGELOG.md for the full release history.
- Q1 2026: Comprehensive documentation ✅
- Q2 2026: TurboQuant KV-cache compression ✅ (shipped in 0.18.0); Qwen/LLaMA tokenizers ✅ (shipped in 0.20.0)
- Q3 2026: Agentic AI enhancements ✅ (tool calling shipped in 0.13.0; ongoing)
- Q4 2026: Federated learning support for multi-device training
We love contributions! Whether it's a new operator, documentation, or a bug fix:
- Read our Contribution Guide.
- Check the Good First Issues.
- Open a discussion or issue on GitHub.
Browse the full codebase documentation on DeepWiki.
- Dhia Chemingui (@dhiaspaner) — Android KMP plugin migration (#385, #386)
MIT — see LICENCE.