OptiRS WASM - WebAssembly Bindings for OptiRS

Version: 0.3.1
Status: Production Ready

High-performance WebAssembly bindings for OptiRS deep learning optimizers and learning rate schedulers. Run state-of-the-art ML optimization algorithms in the browser and Node.js.

Features

• 11 Optimizers - SGD, Adam, AdamW, RAdam, LAMB, Lion, LARS, Adagrad, AdaDelta, AdaBound, Ranger
• 12+ Schedulers - CosineAnnealing, WarmRestarts, CyclicLR, ViTLayerDecay, and more
• TypeScript Support - Complete type definitions included
• Multi-Target - Works with bundlers, web, and Node.js
• Zero Dependencies - Pure Rust compiled to WASM

Installation

npm / yarn

npm install @cooljapan/optirs
# or
yarn add @cooljapan/optirs

Browser (ES Module)

<script type="module">
  import init, { WasmAdam, WasmCosineAnnealing } from '@cooljapan/optirs';
  await init();
  
  const optimizer = WasmAdam.new(0.001);
</script>

Quick Start

Basic Optimization

import init, { WasmAdam, WasmCosineAnnealing } from '@cooljapan/optirs';

await init();

// Create optimizer
const adam = WasmAdam.new(0.001);

// Parameters and gradients as Float64Array
const params = new Float64Array([1.0, 2.0, 3.0, 4.0]);
const grads = new Float64Array([0.1, 0.2, 0.15, 0.08]);

// Optimization step
const updatedParams = adam.step(params, grads);
console.log(updatedParams); // Updated parameters

With Learning Rate Scheduler

import init, { WasmAdamW, WasmCosineAnnealingWarmRestarts } from '@cooljapan/optirs';

await init();

const optimizer = WasmAdamW.new_with_weight_decay(0.001, 0.01);
const scheduler = WasmCosineAnnealingWarmRestarts.new(0.001, 10, 2);

for (let epoch = 0; epoch < 100; epoch++) {
  // Training step...
  const lr = scheduler.step();
  optimizer.set_learning_rate(lr);
}

Advanced Configuration

import init, { WasmOptimizerConfig, WasmAdam } from '@cooljapan/optirs';

await init();

const config = WasmOptimizerConfig.new();
config.learning_rate = 0.001;
config.beta1 = 0.9;
config.beta2 = 0.999;
config.epsilon = 1e-8;
config.weight_decay = 0.01;

const optimizer = WasmAdam.from_config(config);

Available Optimizers

| Optimizer | Constructor | Description | |-----------|-------------|-------------| | WasmSgd | new(lr) | Stochastic Gradient Descent | | WasmAdam | new(lr) | Adaptive Moment Estimation | | WasmAdamW | new_with_weight_decay(lr, wd) | Adam with decoupled weight decay | | WasmRAdam | new(lr) | Rectified Adam | | WasmLamb | new(lr) | Layer-wise Adaptive Moments | | WasmLion | new(lr) | Evolved Sign Momentum | | WasmLars | new(lr) | Layer-wise Adaptive Rate Scaling | | WasmAdagrad | new(lr) | Adaptive Gradient | | WasmAdaDelta | new(rho) | Adaptive Delta | | WasmAdaBound | new(lr) | Bounded adaptive learning rates | | WasmRanger | new(lr) | RAdam + Lookahead |

Available Schedulers

| Scheduler | Constructor | Description | |-----------|-------------|-------------| | WasmConstant | new(lr) | Constant learning rate | | WasmLinearDecay | new(start, end, steps) | Linear interpolation | | WasmStepDecay | new(lr, gamma, step_size) | Step-wise decay | | WasmExponentialDecay | new(lr, gamma) | Exponential decay | | WasmCosineAnnealing | new(lr, t_max) | Cosine annealing | | WasmCosineAnnealingWarmRestarts | new(lr, t_0, t_mult) | Cosine with warm restarts | | WasmCyclicLR | new(base, max, step_up) | Cyclic learning rate | | WasmReduceOnPlateau | new(lr, factor, patience) | Reduce on plateau | | WasmViTLayerDecay | new(base, decay, layers) | Vision Transformer layer decay | | WasmAttentionAwareScheduler | new(base) | Transformer component-specific LR |

Metrics Collection

import init, { WasmMetricsCollector } from '@cooljapan/optirs';

await init();

const metrics = WasmMetricsCollector.new();
metrics.record_learning_rate(0.001);
metrics.record_gradient_norm(0.05);

const stats = metrics.get_stats();
console.log(stats);

Build from Source

# Install wasm-pack
cargo install wasm-pack

# Build for bundler (webpack, rollup, etc.)
./build-wasm.sh bundler

# Build for web (ES modules)
./build-wasm.sh web

# Build for Node.js
./build-wasm.sh nodejs

# Build all targets
./build-wasm.sh all

TypeScript Support

Full TypeScript definitions are included. Import types directly:

import type { WasmOptimizerConfig } from '@cooljapan/optirs';

Browser Compatibility

• Chrome 57+
• Firefox 52+
• Safari 11+
• Edge 79+
• Node.js 12+

Performance

OptiRS WASM provides near-native performance for optimization:

• Compiled with optimizations (opt-level = 3, lto = true)
• No JavaScript overhead in core computation
• SIMD support where available

Links

• OptiRS Documentation
• OptiRS Repository
• npm Package

License

Apache-2.0

Copyright (c) 2026 COOLJAPAN OU (Team Kitasan)