skainet

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 {
    implementation("sk.ainet.core:SKaiNET-lang-core:0.18.0")
    implementation("sk.ainet.core:SKaiNET-backend-cpu:0.18.0")
}

Java / Maven users — see Java Getting Started for BOM setup and JVM flags.

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:

• 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. See TurboQuantUsage for integration guide.

• ComputeGraph: Unified framework for defining agentic workflows and tool-calling loops.
• Java facade: JavaAgentLoop (in skainet-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

• SKaiNET entry point, TensorJavaOps, builder-pattern model definition
• Maven BOM (sk.ainet:skainet-bom) for one-line version management
• Docs: Getting Started | Model Training

• Export trained models to standalone, optimized C99 with static memory allocation
• Ready-to-use Arduino library output
• See arduino-c-codegen.md

• 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
• See hlo-getting-started.md

• TurboQuant KV-Cache Compression — Runtime KV-cache compression for LLM inference: ~8x memory reduction with 4-bit, works with any model (LLaMA, Mistral, Gemma, Qwen). One-line integration via KvCacheStore.turboQuant("balanced", ...).
• Memory Architecture Hardening — First-class storage/placement abstractions (TensorStorage, TensorEncoding, BufferHandle, Placement), zero-copy ownership semantics, quantization-preserving loaders.
• KV-Cache Subsystem — Dedicated KvCacheStore with append-by-token writes, layer/head addressing, asymmetric K/V encoding policies, and CompressedKvAttention SDPA bridge.
• Mistral Tokenizer — Tekken (tiktoken-based BPE) tokenizer support for Mistral models.
• Large Tensor Fix — Fixed Int overflow in GGUF and SafeTensors loaders for tensors > 2 GB (Gemma 4 E4B support).
• CPU SIMD Kernels — Java Vector API acceleration for TurboQuant encode/decode/rotation operations.

See CHANGELOG.md for the full release history.

• Q1 2026: Comprehensive documentation ✅
• Q2 2026: TurboQuant KV-cache compression ✅ (shipped in 0.18.0)
• 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:

1. Read our Contribution Guide.
2. Check the Good First Issues.
3. 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.