> ## Documentation Index
> Fetch the complete documentation index at: https://kumo.ai/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# weight_decay

#### `weight_decay: <list[float]>` (Optional)

## Description

A list of potential `weight_decay` values for AutoML to explore.

Weight decay, also known as L2 regularization, is a technique that penalizes large weights in neural networks. It encourages smaller, more stable weight values, which can help the model generalize better and avoid overfitting.

Each value in `weight_decay` must be `>= 0.0`.

### Supported Task Types

* All

### Default Values

| run\_mode | Default Value             |
| --------- | ------------------------- |
| FAST      | `[0.0, 5e-8, 5e-7, 5e-6]` |
| NORMAL    | `[0.0, 5e-8, 5e-7, 5e-6]` |
| BEST      | `[0.0, 5e-8, 5e-7, 5e-6]` |