Yoga Layout WASM v20251128-0d7c580

Yoga Layout v3.1.0 compiled to WASM for flexbox layout calculations.

Included Files

• yoga-v20251128-0d7c580.wasm - Yoga Layout WASM binary
• yoga-v20251128-0d7c580.mjs - ES Module JavaScript bindings
• yoga-sync-v20251128-0d7c580.js - Synchronous JavaScript wrapper
• checksums.txt - SHA256 checksums for verification
• checksums.txt.asc - GPG signature (if available)

Download URLs

https://github.com/SocketDev/socket-btm/releases/download/yoga-layout-v20251128-0d7c580/yoga-v20251128-0d7c580.wasm
https://github.com/SocketDev/socket-btm/releases/download/yoga-layout-v20251128-0d7c580/yoga-v20251128-0d7c580.mjs
https://github.com/SocketDev/socket-btm/releases/download/yoga-layout-v20251128-0d7c580/yoga-sync-v20251128-0d7c580.js

Verification

# Verify checksums
shasum -a 256 -c checksums.txt

# Verify GPG signature (if GPG key is available)
gpg --verify checksums.txt.asc checksums.txt

Usage

// ES Module (async)
import Yoga from './yoga-v20251128-0d7c580.mjs';

// Synchronous wrapper
import Yoga from './yoga-sync-v20251128-0d7c580.js';

Built from Yoga Layout v3.1.0

ONNX Runtime WASM v20251128-0d7c580

ONNX Runtime v1.20.1 compiled to WASM with SIMD and threading support.

Included Files

• ort.wasm - Main WASM binary with SIMD + threading
• ort.mjs - ES module loader
• checksums.txt - SHA256 checksums for verification

Usage

import * as ort from './ort.mjs';

Built from ONNX Runtime v1.20.1

node-smol-builder 20251128-0d7c580

Minimal Node.js v24.10.0 binaries for all platforms.

Platforms

• macOS: arm64, x64
• Linux (glibc): x64, arm64
• Linux (musl/Alpine): x64, arm64
• Windows: x64, arm64

Files

• node-darwin-arm64 - macOS Apple Silicon
• node-darwin-x64 - macOS Intel
• node-linux-x64 - Linux x64 (glibc)
• node-linux-arm64 - Linux ARM64 (glibc)
• node-linux-x64-musl - Alpine Linux x64
• node-linux-arm64-musl - Alpine Linux ARM64
• node-win-x64.exe - Windows x64
• node-win-arm64.exe - Windows ARM64 (cross-compiled)
• checksums.txt - SHA256 checksums

Usage

Download the appropriate binary for your platform and run it:

./node-darwin-arm64 script.js

Built from Node.js v24.10.0

AI Models v20251128-0d7c580

Production AI models for Socket BTM, optimized with INT4 quantization.

Included Models

MiniLM-L6-v2

• Sentence embeddings model
• 384-dimensional embeddings
• INT4 quantized (~75% size reduction)

CodeT5

• Code understanding model
• INT4 quantized (~75% size reduction)

Files

• models-v20251128-0d7c580.tar.gz - All production models
• checksums.txt - SHA256 checksums for all .onnx files
• checksums.txt.asc - GPG signature (if available)

Download URL

https://github.com/SocketDev/socket-btm/releases/download/models-v20251128-0d7c580/models-v20251128-0d7c580.tar.gz

Verification

# Extract and verify checksums
tar -xzf models-v20251128-0d7c580.tar.gz
shasum -a 256 -c checksums.txt

# Verify GPG signature (if GPG key is available)
gpg --verify checksums.txt.asc checksums.txt

Usage

Extract the archive and load models with ONNX Runtime:

import * as ort from 'onnxruntime-node';
const session = await ort.InferenceSession.create('./minilm-l6/model.onnx');