README.md

yscv-optim

Optimizers, learning rate schedulers, and gradient clipping for neural network training.

use yscv_optim::*;

let mut optimizer = Adam::new(parameters, 1e-3);
let scheduler = CosineScheduler::new(100, 1e-3, 1e-6);

for epoch in 0..100 {
    optimizer.set_lr(scheduler.get_lr(epoch));
    optimizer.step();
    optimizer.zero_grad();
}

Optimizers (8 + Lookahead meta-optimizer)

Sgd, Adam, AdamW, RmsProp, RAdam, Lars, Lamb, Adagrad, plus Lookahead<O> which wraps any of them.

LR Schedulers (11)

StepLr, MultiStepLr, ExponentialLr, CosineAnnealingLr, CosineAnnealingWarmRestarts, LinearWarmupLr, PolynomialDecayLr, OneCycleLr, ReduceLrOnPlateau, CyclicLr, LambdaLr.

Gradient Clipping

• clip_grad_norm — L2 norm clipping
• clip_grad_value — element-wise value clipping

Tests

76 tests covering optimizer convergence, scheduler curves, clipping behavior.