### Initialize Muon Optimizer with Default Settings

Source: https://burn.dev/docs/burn/optim/struct.MuonConfig.html

Demonstrates how to initialize the Muon optimizer using its default configuration. This is the most basic way to get started with the optimizer.

```rust
use burn_optim::{MuonConfig, AdjustLrFn};

// Using default (Original) method
let optimizer = MuonConfig::new().init();
```

--------------------------------

### Example Searches

Source: https://burn.dev/docs/burn/nn/struct.LocalResponseNorm.html?search=

These are example searches that demonstrate how to query LocalResponseNorm.

```APIDOC
## Example Searches

### Description
Demonstrates various ways to search within the LocalResponseNorm functionality.

### Searches
- `std::vec`
- `u32 -> bool`
- `Option<T>, (T -> U) -> Option<U>`
```

--------------------------------

### BinFileRecorder Search Examples

Source: https://burn.dev/docs/burn/record/struct.BinFileRecorder.html?search=

Examples demonstrating how to search within BinFileRecorder.

```APIDOC
## BinFileRecorder Search

This section provides examples of how to search within the BinFileRecorder.

### Example Searches

- `std::vec`
- `u32 -> bool`
- `Option<T>, (T -> U) -> Option<U>`
```

--------------------------------

### CPU Temperature Search Examples

Source: https://burn.dev/docs/burn/train/metric/struct.CpuTemperature.html?search=

This section provides examples of how to search for CPU temperature metrics. These examples demonstrate common search patterns and can be adapted for specific needs.

```APIDOC
## CPU Temperature Search

### Description
Provides examples for searching CPU temperature metrics.

### Examples

#### Basic Search
```
std::vec
```

#### Type Conversion Search
```
u32 -> bool
```

#### Generic Type Search
```
Option<T>, (T -> U) -> Option<U>
```
```

--------------------------------

### ActivationRecordItem Search Examples

Source: https://burn.dev/docs/burn/nn/activation/enum.ActivationRecordItem.html?search=

Provides examples of how to search for ActivationRecordItem. These examples demonstrate common search patterns and type conversions.

```APIDOC
## Search Examples

This section provides examples of how to search for `ActivationRecordItem`.

### Example 1: Searching for a specific type

```rust
// Example search for std::vec
// search("std::vec")
```

### Example 2: Searching for type conversions

```rust
// Example search for u32 to bool conversion
// search("u32 -> bool")
```

### Example 3: Searching for generic type conversions

```rust
// Example search for Option<T>, (T -> U) -> Option<U>
// search("Option<T>, (T -> U) -> Option<U>")
```
```

--------------------------------

### CpuUse Search Examples

Source: https://burn.dev/docs/burn/train/metric/struct.CpuUse.html?search=

Examples of how to search for CPU usage metrics.

```APIDOC
## CpuUse Search

### Description
Provides methods to search and filter CPU usage metrics.

### Example Searches

- `std::vec`
- `u32 -> bool`
- `Option<T>, (T -> U) -> Option<U>`
```

--------------------------------

### Slice Search Examples

Source: https://burn.dev/docs/burn/tensor/struct.Slice.html?search=

Examples demonstrating how to use slice search.

```APIDOC
## Slice Search

### Description
Provides examples for slice searching.

### Example Searches

- `std::vec`
- `u32 -> bool`
- `Option<T>, (T -> U) -> Option<U>`
```

--------------------------------

### AsyncCheckpointer Search Examples

Source: https://burn.dev/docs/burn/train/checkpoint/struct.AsyncCheckpointer.html?search=

Examples demonstrating how to search for different types and conversions within the AsyncCheckpointer system.

```APIDOC
## AsyncCheckpointer Search

This section provides examples of how to use the search functionality for AsyncCheckpointer.

### Example Searches

- `std::vec`
- `u32 -> bool`
- `Option<T>, (T -> U) -> Option<U>`
```

--------------------------------

### FullPrecisionSettings Search Examples

Source: https://burn.dev/docs/burn/record/struct.FullPrecisionSettings.html?search=

Examples of how to search for FullPrecisionSettings.

```APIDOC
## Search Examples

### Example 1: Searching for a specific type

```rust
// Example: Searching for std::vec
search("std::vec");
```

### Example 2: Searching for a type conversion

```rust
// Example: Searching for u32 -> bool conversion
search("u32 -> bool");
```

### Example 3: Searching for a generic type conversion

```rust
// Example: Searching for Option<T>, (T -> U) -> Option<U>
search("Option<T>, (T -> U) -> Option<U>");
```
```

--------------------------------

### EventStoreClient Search Examples

Source: https://burn.dev/docs/burn/train/metric/store/struct.EventStoreClient.html?search=

Provides examples of how to construct search queries for the EventStoreClient.

```APIDOC
## EventStoreClient Search

### Description
This section provides examples of search queries that can be used with the EventStoreClient.

### Example Searches

- `std::vec`
- `u32 -> bool`
- `Option<T>, (T -> U) -> Option<U>`
```

--------------------------------

### Slice Construction Examples

Source: https://burn.dev/docs/burn/tensor/struct.Slice.html

Provides examples of how to construct Slice objects directly.

```APIDOC
## Examples

While you typically use the `s!` macro, you can also construct slices directly:

```rust
use burn_tensor::Slice;

// Equivalent to s![2..8]
let slice1 = Slice::new(2, Some(8), 1);

// Equivalent to s![0..10;2]
let slice2 = Slice::new(0, Some(10), 2);

// Equivalent to s![..;-1] (reverse)
let slice3 = Slice::new(0, None, -1);
```
```

--------------------------------

### Search Examples

Source: https://burn.dev/docs/burn/record/struct.NamedMpkFileRecorder.html?search=

Provides examples of how to formulate search queries for NamedMpkFileRecorder. These examples demonstrate searching for specific types, type conversions, and generic type manipulations.

```APIDOC
## Search Examples

### Description

This section provides examples of search queries that can be used with NamedMpkFileRecorder. These examples illustrate common search patterns and type manipulations.

### Example Searches

- `std::vec`
- `u32 -> bool`
- `Option<T>, (T -> U) -> Option<U>`
```

--------------------------------

### AgentEnvAsyncLoop Search Examples

Source: https://burn.dev/docs/burn/train/struct.AgentEnvAsyncLoop.html?search=

Examples demonstrating how to use the search functionality for AgentEnvAsyncLoop.

```APIDOC
## Search Examples

### Example 1: Searching for a standard type
```
std::vec
```

### Example 2: Searching for a type conversion (integer to boolean)
```
u32 -> bool
```

### Example 3: Searching for a generic type conversion with Option
```
Option<T>, (T -> U) -> Option<U>
```
```

--------------------------------

### AsyncProcessorEvaluation Search Examples

Source: https://burn.dev/docs/burn/train/struct.AsyncProcessorEvaluation.html?search=

Provides examples of search queries that can be performed.

```APIDOC
## Search Examples

### Example 1
```
std::vec
```

### Example 2
```
u32 -> bool
```

### Example 3
```
Option<T>, (T -> U) -> Option<U>
```
```

--------------------------------

### Grid Options Search Examples

Source: https://burn.dev/docs/burn/tensor/grid/struct.GridOptions.html?search=

Examples demonstrating how to search for different types and conversions within the Grid Options.

```APIDOC
## Grid Options Search

This section provides examples of search queries for the Grid Options module.

### Example Searches

- `std::vec`
- `u32 -> bool`
- `Option<T>, (T -> U) -> Option<U>`
```

--------------------------------

### Example Search Queries

Source: https://burn.dev/docs/burn/grad_clipping/enum.GradientClippingConfig.html?search=std%3A%3Avec

Demonstrates example search queries for finding Rust types and functionalities. These are useful for navigating documentation or codebases.

```text
Example searches:
  * std::vec
  * u32 -> bool
  * Option<T>, (T -> U) -> Option<U>
```

--------------------------------

### Metric Definition Search Examples

Source: https://burn.dev/docs/burn/train/metric/struct.MetricDefinition.html?search=

Examples of how to search for metric definitions. These examples demonstrate different query patterns.

```APIDOC
## Search Examples

### Example 1: Searching for a specific type

```
std::vec
```

### Example 2: Searching for a type conversion

```
u32 -> bool
```

### Example 3: Searching for a generic function signature

```
Option<T>, (T -> U) -> Option<U>
```
```

--------------------------------

### BinBytesRecorder Search Examples

Source: https://burn.dev/docs/burn/record/struct.BinBytesRecorder.html?search=

Provides examples of search queries that can be performed on BinBytesRecorder.

```APIDOC
## Search Examples

### Example Queries

- `std::vec`
- `u32 -> bool`
- `Option<T>, (T -> U) -> Option<U>`
```

--------------------------------

### EpochSummary Search Examples

Source: https://burn.dev/docs/burn/train/metric/store/struct.EpochSummary.html?search=

Examples of how to search for EpochSummary.

```APIDOC
## EpochSummary Search

### Description
Provides examples of how to search for EpochSummary.

### Example Searches
- `std::vec`
- `u32 -> bool`
- `Option<T>, (T -> U) -> Option<U>`
```

--------------------------------

### RLTraining::new

Source: https://burn.dev/docs/burn/train/struct.RLTraining.html?search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

Creates a new RLTraining instance with a specified directory for checkpoints and an environment initializer.

```APIDOC
## RLTraining::new

### Description
Creates a new runner for reinforcement learning.

### Method
`new(directory: impl AsRef<Path>, env_initializer: EI)`

### Parameters
#### Arguments
* `directory` - The directory to save the checkpoints.
* `env_init` - Specifies how to initialize the environment.
```

--------------------------------

### Split Off Slice Starting with Third Element

Source: https://burn.dev/docs/burn/prelude/struct.Shape.html

Use `split_off` to remove and return a subslice. This example shows splitting off the elements starting from the third element.

```rust
let mut slice: &[_] = &['a', 'b', 'c', 'd'];
let mut tail = slice.split_off(2..).unwrap();

assert_eq!(slice, &['a', 'b']);
assert_eq!(tail, &['c', 'd']);
```

--------------------------------

### LpLossConfig Examples

Source: https://burn.dev/docs/burn/nn/loss/struct.LpLossConfig.html

Demonstrates how to create L1, L2, and custom Lp loss configurations using LpLossConfig.

```rust
use burn_nn::loss::{LpLossConfig, Reduction};

// Create L1 loss (MAE when using mean reduction)
let l1_loss = LpLossConfig::l1();

// Create L2 loss (MSE when using mean reduction)
let l2_loss = LpLossConfig::l2();

// Create custom Lp loss with p=3
let l3_loss = LpLossConfig::new(3.0).init();
```

--------------------------------

### Common Search Examples

Source: https://burn.dev/docs/burn/nn/modules/norm/struct.GroupNormConfig.html?search=

Illustrates typical search queries for various data types and generic structures. These examples can guide users on how to formulate their own searches.

```rust
* std::vec
```

```rust
* u32 -> bool
```

```rust
* Option<T>, (T -> U) -> Option<U>
```

--------------------------------

### LpLossConfig Initialization

Source: https://burn.dev/docs/burn/nn/loss/struct.LpLossConfig.html?search=

Demonstrates how to create LpLossConfig instances for L1, L2, and custom Lp losses.

```APIDOC
## LpLossConfig
burn::nn::loss

Configuration for the Lp Loss module.

### §Example
```
use burn_nn::loss::{LpLossConfig, Reduction};

// Create L1 loss (MAE when using mean reduction)
let l1_loss = LpLossConfig::l1();

// Create L2 loss (MSE when using mean reduction)
let l2_loss = LpLossConfig::l2();

// Create custom Lp loss with p=3
let l3_loss = LpLossConfig::new(3.0).init();
```

## Fields
§`p: f64`
The exponent `p` determining the type of error measurement.
Common values:
  * `p = 1.0`: L1 loss (MAE with mean reduction) - robust to outliers
  * `p = 2.0`: L2 loss (MSE with mean reduction) - standard choice, differentiable everywhere
  * `p > 2.0`: Increasingly sensitive to large errors (outliers)
  * `0 < p < 1`: More robust to outliers than L1 (quasi-norm)

## Implementations
### impl LpLossConfig
#### pub fn new(p: f64) -> LpLossConfig
Create a new instance of the config.
##### §Arguments
###### §Required Arguments
###### §`p`
The exponent `p` determining the type of error measurement.
Common values:
  * `p = 1.0`: L1 loss (MAE with mean reduction) - robust to outliers
  * `p = 2.0`: L2 loss (MSE with mean reduction) - standard choice, differentiable everywhere
  * `p > 2.0`: Increasingly sensitive to large errors (outliers)
  * `0 < p < 1`: More robust to outliers than L1 (quasi-norm)

#### pub fn init(&self) -> LpLoss
Initializes a Lp Loss module.
##### §Panics
Panics if `p <= 0`.
#### pub fn l1() -> LpLoss
Creates L1 loss (p=1).
When used with `Reduction::Mean`, this computes Mean Absolute Error (MAE). When used with `Reduction::Sum`, this computes Sum of Absolute Errors (SAE).
#### pub fn l2() -> LpLoss
Creates L2 loss (p=2).
When used with `Reduction::Mean`, this computes Mean Squared Error (MSE). When used with `Reduction::Sum`, this computes Sum of Squared Errors (SSE).
```

--------------------------------

### Split Off Mutable Slice Starting with Third Element

Source: https://burn.dev/docs/burn/prelude/struct.Shape.html

Use `split_off_mut` to remove and return a mutable subslice. This example shows splitting off the elements starting from the third element.

```rust
let mut slice: &mut [_] = &mut ['a', 'b', 'c', 'd'];
let mut tail = slice.split_off_mut(2..).unwrap();

assert_eq!(slice, &mut ['a', 'b']);
assert_eq!(tail, &mut ['c', 'd']);
```

--------------------------------

### LearnerSummaryConfig Search Examples

Source: https://burn.dev/docs/burn/train/struct.LearnerSummaryConfig.html?search=

Examples of how to perform searches on LearnerSummaryConfig.

```APIDOC
## LearnerSummaryConfig Search

### Description
Provides examples for searching within LearnerSummaryConfig.

### Example Searches

- `std::vec`
- `u32 -> bool`
- `Option<T>, (T -> U) -> Option<U>`
```

--------------------------------

### Example Initializer Searches

Source: https://burn.dev/docs/burn/module/enum.Initializer.html?search=

Demonstrates various ways to search for initializers, including searching by type, type transformations, and generic types.

```APIDOC
## Example Searches

### Description
Provides examples of how to search for initializers.

### Code Examples

#### Search by Type
```apidoc
std::vec
```

#### Search by Type Transformation
```apidoc
u32 -> bool
```

#### Search with Generic Types and Transformations
```apidoc
Option<T>, (T -> U) -> Option<U>
```
```

--------------------------------

### ActivationConfig_search

Source: https://burn.dev/docs/burn/nn/activation/enum.ActivationConfig.html?search=

This section details how to search for activation configurations. It includes example searches to guide users.

```APIDOC
## ActivationConfig_search

### Description
Provides functionality to search for activation configurations.

### Example Searches
- `std::vec`
- `u32 -> bool`
- `Option<T>, (T -> U) -> Option<U>`
```

--------------------------------

### PrecisionMetric Search Examples

Source: https://burn.dev/docs/burn/train/metric/struct.PrecisionMetric.html?search=

Examples of how to search for PrecisionMetric related items.

```APIDOC
## PrecisionMetric Search

### Description
Provides examples for searching PrecisionMetric related functionalities.

### Example Searches
* `std::vec`
* `u32 -> bool`
* `Option<T>, (T -> U) -> Option<U>`
```

--------------------------------

### install

Source: https://burn.dev/docs/burn/train/trait.ApplicationLoggerInstaller.html

Installs the application logger. Returns a Result indicating success or failure with an error message.

```APIDOC
## fn install(&self) -> Result<(), String>

### Description
Install the application logger.

### Returns
- `Result<(), String>`: Ok(()) on success, or an Err(String) with an error message on failure.
```

--------------------------------

### Get Window by Index

Source: https://burn.dev/docs/burn/data/dataset/transform/struct.WindowsDataset.html?search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

Retrieves a specific window of data as a vector. The index corresponds to the starting position of the window in the dataset.

```rust
fn get(&self, index: usize) -> Option<Vec<I>>

```

--------------------------------

### EluConfig Initialization and Configuration

Source: https://burn.dev/docs/burn/nn/activation/struct.EluConfig.html?search=

Demonstrates how to create and configure an EluConfig object.

```APIDOC
## Struct EluConfig

Configuration to create an Elu layer using the init function.

### Fields

*   `alpha`: `f64` - The alpha value. Default is 1.0

### Methods

*   `pub fn new() -> EluConfig`
    Create a new instance of the config with default alpha value.

*   `pub fn with_alpha(self, alpha: f64) -> EluConfig`
    Sets the value for the `alpha` field.

*   `pub fn init(&self) -> Elu`
    Initialize a new Elu Layer using the current configuration.
```

--------------------------------

### LocalResponseNorm Search Examples

Source: https://burn.dev/docs/burn/nn/modules/struct.LocalResponseNorm.html?search=

Provides examples of how to perform searches within the LocalResponseNorm module.

```APIDOC
## Search Examples

### Description
Examples of common search queries.

### Example Searches
* `std::vec`
* `u32 -> bool`
* `Option<T>, (T -> U) -> Option<U>`
```

--------------------------------

### InstanceNormRecordItem Search Examples

Source: https://burn.dev/docs/burn/nn/struct.InstanceNormRecordItem.html?search=

Examples of how to search for InstanceNormRecordItem, including common patterns and types.

```APIDOC
## InstanceNormRecordItem Search

### Description
Provides examples of search queries for InstanceNormRecordItem.

### Example Searches
- `std::vec`
- `u32 -> bool`
- `Option<T>, (T -> U) -> Option<U>`
```

--------------------------------

### CTCLoss Initialization and Usage Example

Source: https://burn.dev/docs/burn/nn/loss/struct.CTCLoss.html?search=

Demonstrates how to initialize CTCLoss with default and custom configurations, and how to compute the loss using log-probabilities, targets, and lengths.

```rust
use burn::tensor::{Tensor, Int};
use burn::tensor::activation::log_softmax;
use burn::nn::loss::{CTCLossConfig, CTCLoss};

let device = Default::default();

// Initialize CTC Loss with default configuration
let ctc_loss = CTCLossConfig::new().init();

// Initialize CTC Loss with custom configuration
let ctc_loss = CTCLossConfig::new()
    .with_blank(1)
    .with_zero_infinity(true)
    .init();

// Prepare inputs (Logits shape: [Time, Batch, Class])
// In your actual code, the logits would be the output of your model
let logits = Tensor::<B, 3>::ones([10, 2, 5], &device);
let log_probs = log_softmax(logits, 2);

// Targets shape: [Batch, Max_Target_Len]
// Note: Targets should not contain the blank index (1).
let targets = Tensor::<B, 2, Int>::from_data([[0, 2], [3, 4]], &device);

// Lengths shape: [Batch]
let input_lengths = Tensor::<B, 1, Int>::from_data([10, 8], &device);
let target_lengths = Tensor::<B, 1, Int>::from_data([2, 2], &device);

// Compute loss
let loss = ctc_loss.forward(log_probs, targets, input_lengths, target_lengths);
```

--------------------------------

### SwiGluConfig::init Method

Source: https://burn.dev/docs/burn/nn/activation/struct.SwiGluConfig.html?search=

Initializes a SwiGlu activation layer using the current configuration and a given device.

```rust
pub fn init<B>(&self, device: &<B as BackendTypes>::Device) -> SwiGlu<B>
where B: Backend
```

--------------------------------

### Reverse split subslices

Source: https://burn.dev/docs/burn/prelude/struct.Shape.html

Use `rsplit` to get an iterator over subslices separated by elements matching a predicate, starting from the end of the slice. The matched element is excluded.

```rust
let slice = [11, 22, 33, 0, 44, 55];
let mut iter = slice.rsplit(|num| *num == 0);

assert_eq!(iter.next().unwrap(), &[44, 55]);
assert_eq!(iter.next().unwrap(), &[11, 22, 33]);
assert_eq!(iter.next(), None);
```

```rust
let v = &[0, 1, 1, 2, 3, 5, 8];
let mut it = v.rsplit(|n| *n % 2 == 0);
assert_eq!(it.next().unwrap(), &[]);
assert_eq!(it.next().unwrap(), &[3, 5]);
assert_eq!(it.next().unwrap(), &[1, 1]);
assert_eq!(it.next().unwrap(), &[]);
assert_eq!(it.next(), None);
```

--------------------------------

### Dataset Windowing Example

Source: https://burn.dev/docs/burn/data/dataset/audio/struct.SpeechCommandsDataset.html?search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

Demonstrates how to create sliding windows over a dataset. Panics if the window size is 0.

```rust
use crate::burn_dataset::
   transform::{Windows, WindowsDataset},
   Dataset, InMemDataset,
};

let items = [1, 2, 3, 4].to_vec();
let dataset = InMemDataset::new(items.clone());

for window in dataset.windows(2) {
 // do sth with window
}
```

--------------------------------

### Extending Vec from a Range (Start)

Source: https://burn.dev/docs/burn/module/type.Devices.html

Given a range `src`, clones a slice of elements in that range and appends it to the end of the Vec. This example extends from the beginning up to (but not including) the second element.

```rust
let mut numbers = vec![0, 1, 2, 3, 4];
numbers.extend_from_within(..2);
assert_eq!(numbers, [0, 1, 2, 3, 4, 0, 1]);

```

--------------------------------

### InstanceNormConfig::init

Source: https://burn.dev/docs/burn/nn/modules/norm/struct.InstanceNormConfig.html?search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

Initializes a new InstanceNorm module on the specified device.

```APIDOC
## InstanceNormConfig::init

### Description
Initialize a new instance norm module.

### Signature
```rust
pub fn init<B>(&self, device: &<B as BackendTypes>::Device) -> InstanceNorm<B>
where B: Backend,
```

### Arguments
* **device** (&`<B as BackendTypes>::Device`) - The device on which to initialize the module.
```

--------------------------------

### Reverse split mutable subslices

Source: https://burn.dev/docs/burn/prelude/struct.Shape.html

Use `rsplit_mut` to get an iterator over mutable subslices separated by elements matching a predicate, starting from the end of the slice. The matched element is excluded.

```rust
let mut v = [100, 400, 300, 200, 600, 500];

let mut count = 0;
for group in v.rsplit_mut(|num| *num % 3 == 0) {
    count += 1;
    group[0] = count;
}
assert_eq!(v, [3, 400, 300, 2, 600, 1]);
```

--------------------------------

### InstanceNormConfig::init

Source: https://burn.dev/docs/burn/nn/modules/norm/struct.InstanceNormConfig.html?search=u32+-%3E+bool

Initializes a new InstanceNorm module using the provided configuration and device.

```APIDOC
## InstanceNormConfig::init

### Description
Initialize a new instance norm module.

### Signature
```rust
pub fn init<B: Backend>(&self, device: &<B as BackendTypes>::Device) -> InstanceNorm<B>
```

### Arguments
* `device` (&<B as BackendTypes>::Device) - The device on which to initialize the InstanceNorm module.
```

--------------------------------

### Getting Metadata with `and_then`

Source: https://burn.dev/docs/burn/data/dataset/type.Result.html?search=u32+-%3E+bool

Demonstrates using `and_then` to chain metadata retrieval operations on file paths. This example shows how to handle potential errors during file system access.

```rust
use std::{io::ErrorKind, path::Path};

// Note: on Windows "/" maps to "C:\\"
let root_modified_time = Path::new("/").metadata().and_then(|md| md.modified());
assert!(root_modified_time.is_ok());

let should_fail = Path::new("/bad/path").metadata().and_then(|md| md.modified());
assert!(should_fail.is_err());
assert_eq!(should_fail.unwrap_err().kind(), ErrorKind::NotFound);
```

--------------------------------

### Create and Use WindowsDataset

Source: https://burn.dev/docs/burn/data/dataset/transform/struct.SamplerDataset.html?search=

Demonstrates how to create a dataset and then use the `windows` method to create sliding windows of items. Panics if the window size is 0.

```rust
use crate::burn_dataset::
   transform::{Windows, WindowsDataset},
   Dataset, InMemDataset,
};

let items = [1, 2, 3, 4].to_vec();
let dataset = InMemDataset::new(items.clone());

let windows_dataset = WindowsDataset::new(dataset, 2);
let mut iter = windows_dataset.iter();

assert_eq!(iter.next().unwrap(), vec![1, 2]);
assert_eq!(iter.next().unwrap(), vec![2, 3]);
assert_eq!(iter.next().unwrap(), vec![3, 4]);
assert_eq!(iter.next(), None);

```

--------------------------------

### Mutably Iterate Over Chunks From the End

Source: https://burn.dev/docs/burn/prelude/struct.Shape.html

Use `rchunks_mut` to get mutable references to overlapping chunks of a specified size, starting from the end. The last chunk may be smaller. Panics if `chunk_size` is zero.

```rust
let v = &mut [0, 0, 0, 0, 0];
let mut count = 1;

for chunk in v.rchunks_mut(2) {
    for elem in chunk.iter_mut() {
        *elem += count;
    }
    count += 1;
}
assert_eq!(v, &[3, 2, 2, 1, 1]);
```

--------------------------------

### Search Examples

Source: https://burn.dev/docs/burn/train/struct.AsyncProcessorTraining.html?search=

Examples demonstrating how to use the search functionality to find specific types and patterns.

```APIDOC
## Search Examples

### Description
Examples demonstrating how to use the search functionality to find specific types and patterns.

### Example 1: Searching for a specific type
```
std::vec
```

### Example 2: Searching for a type conversion
```
u32 -> bool
```

### Example 3: Searching for a generic type conversion
```
Option<T>, (T -> U) -> Option<U>
```
```

--------------------------------

### SmoothL1LossConfig::init() Method

Source: https://burn.dev/docs/burn/nn/loss/struct.SmoothL1LossConfig.html?search=u32+-%3E+bool

Initializes a Smooth L1 Loss module using the current configuration. Panics if beta is not positive.

```rust
pub fn init(&self) -> SmoothL1Loss
```

--------------------------------

### Get Pointer Range of Slice

Source: https://burn.dev/docs/burn/prelude/struct.Shape.html

Retrieve the start and end raw pointers of a slice, forming a half-open range. The end pointer is one past the last element and requires extra caution.

```rust
let a = [1, 2, 3];
let x = &a[1] as *const _;
let y = &5 as *const _;

assert!(a.as_ptr_range().contains(&x));
assert!(!a.as_ptr_range().contains(&y));
```

--------------------------------

### LpLossConfig::init Method

Source: https://burn.dev/docs/burn/nn/loss/struct.LpLossConfig.html?search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

Initializes a Lp Loss module from its configuration. Panics if p <= 0.

```rust
pub fn init(&self) -> LpLoss
```

--------------------------------

### Get Mutable Pointer Range of Slice

Source: https://burn.dev/docs/burn/prelude/struct.Shape.html

Obtain the start and end mutable raw pointers of a slice, forming a half-open range. The end pointer is one past the last element and requires extra caution.

```rust
let mut v = [1, 2, 3];
let ptr_range = v.as_mut_ptr_range();
// Example usage would go here, but is not provided in the source.
```

--------------------------------

### LpLossConfig Initialization

Source: https://burn.dev/docs/burn/nn/loss/struct.LpLossConfig.html

Demonstrates how to create LpLossConfig instances for L1, L2, and custom p-values.

```APIDOC
## LpLossConfig

Configuration for the Lp Loss module.

### §Example
```rust
use burn_nn::loss::{LpLossConfig, Reduction};

// Create L1 loss (MAE when using mean reduction)
let l1_loss = LpLossConfig::l1();

// Create L2 loss (MSE when using mean reduction)
let l2_loss = LpLossConfig::l2();

// Create custom Lp loss with p=3
let l3_loss = LpLossConfig::new(3.0).init();
```

## Fields

### `p: f64`
The exponent `p` determining the type of error measurement.
Common values:
  * `p = 1.0`: L1 loss (MAE with mean reduction) - robust to outliers
  * `p = 2.0`: L2 loss (MSE with mean reduction) - standard choice, differentiable everywhere
  * `p > 2.0`: Increasingly sensitive to large errors (outliers)
  * `0 < p < 1`: More robust to outliers than L1 (quasi-norm)

## Implementations

### `impl LpLossConfig`

#### `pub fn new(p: f64) -> LpLossConfig`
Create a new instance of the config.

##### §Arguments
###### §Required Arguments
###### §`p`
The exponent `p` determining the type of error measurement.
Common values:
  * `p = 1.0`: L1 loss (MAE with mean reduction) - robust to outliers
  * `p = 2.0`: L2 loss (MSE with mean reduction) - standard choice, differentiable everywhere
  * `p > 2.0`: Increasingly sensitive to large errors (outliers)
  * `0 < p < 1`: More robust to outliers than L1 (quasi-norm)

#### `pub fn init(&self) -> LpLoss`
Initializes a Lp Loss module.

##### §Panics
Panics if `p <= 0`.

#### `pub fn l1() -> LpLoss`
Creates L1 loss (p=1).
When used with `Reduction::Mean`, this computes Mean Absolute Error (MAE). When used with `Reduction::Sum`, this computes Sum of Absolute Errors (SAE).

#### `pub fn l2() -> LpLoss`
Creates L2 loss (p=2).
When used with `Reduction::Mean`, this computes Mean Squared Error (MSE). When used with `Reduction::Sum`, this computes Sum of Squared Errors (SSE).
```

--------------------------------

### Get Element Offset by Pointer

Source: https://burn.dev/docs/burn/prelude/struct.Shape.html

Use `element_offset` to find the index of an element reference within a slice using pointer arithmetic. It returns `None` if the element reference does not point to the start of an element. This method panics if `T` is zero-sized.

```rust
let nums: &[u32] = &[1, 7, 1, 1];
let num = &nums[2];

assert_eq!(num, &1);
assert_eq!(nums.element_offset(num), Some(2));
```

```rust
let arr: &[[u32; 2]] = &[[0, 1], [2, 3]];
let flat_arr: &[u32] = arr.as_flattened();

let ok_elm: &[u32; 2] = flat_arr[0..2].try_into().unwrap();
let weird_elm: &[u32; 2] = flat_arr[1..3].try_into().unwrap();

assert_eq!(ok_elm, &[0, 1]);
assert_eq!(weird_elm, &[1, 2]);

assert_eq!(arr.element_offset(ok_elm), Some(0)); // Points to element 0
assert_eq!(arr.element_offset(weird_elm), None); // Points between element 0 and 1
```

--------------------------------

### Get Element Offset by Pointer

Source: https://burn.dev/docs/burn/prelude/struct.Shape.html?search=u32+-%3E+bool

Find the index of an element within a slice using its memory address. This method relies on pointer arithmetic and does not compare element values. It returns `None` if the pointer is not aligned to an element's start.

```rust
let nums: &[u32] = &[1, 7, 1, 1];
let num = &nums[2];

assert_eq!(num, &1);
assert_eq!(nums.element_offset(num), Some(2));
```

--------------------------------

### Arc::into_raw_with_allocator and Arc::from_raw_in Example

Source: https://burn.dev/docs/burn/tensor/backend/type.ServerUtilitiesHandle.html?search=

Demonstrates consuming an Arc to get its raw pointer and allocator, and then reconstructing the Arc from these. This requires the `allocator_api` feature and involves unsafe operations for pointer manipulation. Ensure the pointer is converted back to an Arc to avoid memory leaks.

```rust
#![feature(allocator_api)]
use std::sync::Arc;
use std::alloc::System;

let x = Arc::new_in("hello".to_owned(), System);
let (ptr, alloc) = Arc::into_raw_with_allocator(x);
assert_eq!(unsafe { &*ptr }, "hello");
let x = unsafe { Arc::from_raw_in(ptr, alloc) };
assert_eq!(&*x, "hello");
```

--------------------------------

### ViewLayoutLaunchArg Implementations

Source: https://burn.dev/docs/burn/tensor/struct.f16.html?search=u32+-%3E+bool

Handles argument registration and expansion for kernel compilation.

```APIDOC
## type RuntimeArg<R: Runtime> = <T as LaunchArg>::RuntimeArg<R>

### Description
The runtime argument for the kernel.
```

```APIDOC
## type CompilationArg = <T as LaunchArg>::CompilationArg

### Description
Compilation argument.
```

```APIDOC
## fn register<R, B>( arg: <T as ViewLayoutLaunchArg>::RuntimeArg<R>, _buffer: &B, _ty: Type, launcher: &mut KernelLauncher<R>, ) -> <T as ViewLayoutLaunchArg>::CompilationArg

### Description
Registers an input or output argument for a kernel launch.

### Parameters
* `arg`: The runtime argument for the kernel.
* `_buffer`: A reference to the buffer argument.
* `_ty`: The type of the argument.
* `launcher`: A mutable reference to the `KernelLauncher`.

### Returns
The compilation argument for the registered argument.
```

```APIDOC
## fn expand( arg: &<T as ViewLayoutLaunchArg>::CompilationArg, _ty: Type, builder: &mut KernelBuilder, ) -> <T as CubeType>::ExpandType

### Description
Register an input variable during compilation that fills the [`KernelBuilder`].

### Parameters
* `arg`: A reference to the compilation argument.
* `_ty`: The type of the argument.
* `builder`: A mutable reference to the `KernelBuilder`.

### Returns
The expanded type for the kernel builder.
```

```APIDOC
## fn expand_output( arg: &<T as ViewLayoutLaunchArg>::CompilationArg, _ty: Type, builder: &mut KernelBuilder, ) -> <T as CubeType>::ExpandType

### Description
Register an output variable during compilation that fills the [`KernelBuilder`].

### Parameters
* `arg`: A reference to the compilation argument.
* `_ty`: The type of the argument.
* `builder`: A mutable reference to the `KernelBuilder`.

### Returns
The expanded type for the kernel builder.
```

--------------------------------

### RmsPropStateItem Search Examples

Source: https://burn.dev/docs/burn/optim/struct.RmsPropStateItem.html?search=

Examples of how to search for RmsPropStateItem. These examples demonstrate common search patterns.

```APIDOC
## Search Examples

### Example 1: Searching for a specific type
```rust
// Example: Search for std::vec
```

### Example 2: Searching for type conversions
```rust
// Example: Search for u32 -> bool
```

### Example 3: Searching for generic type conversions
```rust
// Example: Search for Option<T>, (T -> U) -> Option<U>
```
```

--------------------------------

### MultiDevicesTrainStep::new

Source: https://burn.dev/docs/burn/train/train/struct.MultiDevicesTrainStep.html?search=

Creates a new instance of MultiDevicesTrainStep, initializing it for multi-device training.

```APIDOC
## MultiDevicesTrainStep::new

### Description
Create a new multi devices train step.

### Method
`pub fn new(devices: &[<<LC as LearningComponentsTypes>::Backend as BackendTypes>::Device]) -> MultiDevicesTrainStep<LC>`

### Arguments
* `devices` - Devices to be used for training.

### Returns
A new `MultiDevicesTrainStep` instance.
```

--------------------------------

### MomentumStateItem Search Examples

Source: https://burn.dev/docs/burn/optim/momentum/struct.MomentumStateItem.html?search=

Examples of how to search for MomentumStateItem. These examples demonstrate common search patterns.

```APIDOC
## Search Examples

### Example 1: Searching for a specific type

```
std::vec
```

### Example 2: Searching for a type conversion

```
u32 -> bool
```

### Example 3: Searching for a generic type conversion

```
Option<T>, (T -> U) -> Option<U>
```
```

--------------------------------

### CrossEntropyLossRecordItem Search Examples

Source: https://burn.dev/docs/burn/nn/loss/struct.CrossEntropyLossRecordItem.html?search=

Examples of how to search for CrossEntropyLossRecordItem. These examples demonstrate different query patterns.

```APIDOC
## Search Examples

### Example 1: Searching for a specific type
```
std::vec
```

### Example 2: Searching for type conversions
```
u32 -> bool
```

### Example 3: Searching for generic type transformations
```
Option<T>, (T -> U) -> Option<U>
```
```

--------------------------------

### InstanceNormConfig::init

Source: https://burn.dev/docs/burn/nn/modules/norm/struct.InstanceNormConfig.html

Initializes a new InstanceNorm module using the configuration.

```APIDOC
## InstanceNormConfig::init

### Description
Initialize a new instance norm module.

### Method
`pub fn init<B>(&self, device: &<B as BackendTypes>::Device) -> InstanceNorm<B>`

### Type Parameters
* `B` - Represents the backend type.

### Arguments
* `device` (&`<B as BackendTypes>::Device`) - The device on which to initialize the InstanceNorm module.
```

--------------------------------

### ComposedLrSchedulerRecordItem Search Examples

Source: https://burn.dev/docs/burn/lr_scheduler/composed/struct.ComposedLrSchedulerRecordItem.html?search=

Examples of how to search for ComposedLrSchedulerRecordItem. These examples demonstrate different search patterns.

```APIDOC
## Search Examples

### Example 1: Searching for a specific item
```
std::vec
```

### Example 2: Searching for a type conversion
```
u32 -> bool
```

### Example 3: Searching for a generic type conversion
```
Option<T>, (T -> U) -> Option<U>
```
```

--------------------------------

### SoftplusConfig init() Method

Source: https://burn.dev/docs/burn/nn/activation/struct.SoftplusConfig.html?search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

Initializes a new Softplus Layer using the current configuration.

```rust
pub fn init(&self) -> Softplus
```

--------------------------------

### MaxPool2dWithIndices Search Examples

Source: https://burn.dev/docs/burn/tensor/backend/ops/struct.MaxPool2dWithIndices.html?search=

Examples of how to search for MaxPool2dWithIndices operations. These examples demonstrate different query patterns.

```APIDOC
## MaxPool2dWithIndices Search

### Description
Provides examples of search queries for `MaxPool2dWithIndices`.

### Example Searches

- `std::vec`
- `u32 -> bool`
- `Option<T>, (T -> U) -> Option<U>`
```

--------------------------------

### Dataset Windows Example

Source: https://burn.dev/docs/burn/data/dataset/transform/struct.SamplerDataset.html

Demonstrates how to create a sliding window over a dataset using the `windows` method. This is useful for processing sequences of data. Ensure the dataset is not empty and the window size is valid.

```rust
use crate::burn_dataset::
   transform::{Windows, WindowsDataset},
   Dataset, InMemDataset,
};

let items = [1, 2, 3, 4].to_vec();
let dataset = InMemDataset::new(items.clone());

let windows_dataset = WindowsDataset::new(dataset, 2);
let windows = windows_dataset.windows(2);

assert_eq!(windows.next().unwrap(), vec![1, 2]);
assert_eq!(windows.next().unwrap(), vec![2, 3]);
assert_eq!(windows.next().unwrap(), vec![3, 4]);
assert!(windows.next().is_none());
```

--------------------------------

### ConvTranspose1dRecordItem Search Examples

Source: https://burn.dev/docs/burn/nn/modules/conv/struct.ConvTranspose1dRecordItem.html?search=

Examples of how to perform searches for ConvTranspose1dRecordItem. These examples demonstrate different query patterns.

```APIDOC
## Search Examples

### Example 1: Basic Type Search

```rust
// Search for std::vec
```

### Example 2: Type Conversion Search

```rust
// Search for u32 -> bool
```

### Example 3: Generic Type Conversion Search

```rust
// Search for Option<T>, (T -> U) -> Option<U>
```
```

--------------------------------

### RngSource Initialization Examples

Source: https://burn.dev/docs/burn/data/dataset/transform/enum.RngSource.html?search=u32+-%3E+bool

Demonstrates various ways to create an RngSource instance, including default, seeded, and from an existing StdRng. Also shows how to fork an RNG.

```rust
use rand::rngs::StdRng;
use rand::SeedableRng;
use burn_dataset::transform::RngSource;

// Default via `StdRng::from_os_rng()` (`RngSource::Default`)
let system: RngSource = RngSource::default();

// From a fixed seed (`RngSource::Seed`)
let seeded: RngSource = 42.into();

// From an existing rng (`RngSource::Rng`)
let rng = StdRng::seed_from_u64(123);
let with_rng: RngSource = rng.into();

// Forks the parent RNG to derive an independent, deterministic child RNG.
// The original `rng` is modified, and the resulting `RngSource` contains
// a new RNG starting from a unique state.
let mut rng = StdRng::seed_from_u64(123);
let forked: RngSource = (&mut rng).into();
```

--------------------------------

### Install Application Logger

Source: https://burn.dev/docs/burn/train/struct.FileApplicationLoggerInstaller.html

Installs the configured file application logger. This method is part of the ApplicationLoggerInstaller trait implementation for FileApplicationLoggerInstaller.

```rust
fn install(&self) -> Result<(), String>
```

--------------------------------

### FileApplicationLoggerInstaller::install

Source: https://burn.dev/docs/burn/train/struct.FileApplicationLoggerInstaller.html?search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

Installs the configured file application logger, making it active for log output.

```APIDOC
## FileApplicationLoggerInstaller::install

### Description
Install the application logger.

### Method

```rust
fn install(&self) -> Result<(), String>
```

### Returns

* `Result<(), String>` - Ok(()) if the logger was installed successfully, or an Err(String) with an error message if installation failed.
```

--------------------------------

### EluConfig::init()

Source: https://burn.dev/docs/burn/nn/activation/struct.EluConfig.html?search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

Initializes a new Elu Layer using the current configuration.

```APIDOC
### `fn init(&self) -> Elu`

Initialize a new Elu Layer.
```

--------------------------------

### BiLstmRecord Search Examples

Source: https://burn.dev/docs/burn/nn/struct.BiLstmRecord.html?search=

Provides examples of how to perform searches within the BiLstmRecord system. These examples illustrate common search patterns and syntax.

```APIDOC
## Search Examples

### General Search Syntax

- **std::vec**: Search for items related to the standard vector type.
- **u32 -> bool**: Search for functions or types that convert a u32 to a boolean.
- **Option<T>, (T -> U) -> Option<U>**: Search for operations involving Option types and transformations.

### Usage

These examples can be used as a starting point for constructing your own search queries to find specific data or functionality within the BiLstmRecord system.
```

--------------------------------

### FileApplicationLoggerInstaller::new

Source: https://burn.dev/docs/burn/train/struct.FileApplicationLoggerInstaller.html

Creates a new instance of FileApplicationLoggerInstaller, configured to log to the specified file path.

```APIDOC
## pub fn new(path: impl AsRef<Path>) -> FileApplicationLoggerInstaller

### Description
Create a new file application logger.

### Parameters
#### Path Parameters
- **path** (impl AsRef<Path>) - Required - The path to the file where logs will be written.
```

--------------------------------

### LstmRecordItem_search Examples

Source: https://burn.dev/docs/burn/nn/modules/lstm/struct.LstmRecordItem.html?search=

Examples of how to search for LstmRecordItem records.

```APIDOC
## Search Examples

This section provides example search queries for `LstmRecordItem`.

### Example Searches

- `std::vec`
- `u32 -> bool`
- `Option<T>, (T -> U) -> Option<U>`
```

--------------------------------

### ComposedLrSchedulerRecord Search Examples

Source: https://burn.dev/docs/burn/lr_scheduler/composed/struct.ComposedLrSchedulerRecord.html?search=

Examples of how to search for ComposedLrSchedulerRecord.

```APIDOC
## Search Examples

Example searches:
  * std::vec
  * u32 -> bool
  * Option<T>, (T -> U) -> Option<U>
```

--------------------------------

### RLTraining::new

Source: https://burn.dev/docs/burn/train/struct.RLTraining.html

Creates a new runner for reinforcement learning with a specified directory for checkpoints and an environment initializer.

```APIDOC
## RLTraining::new

### Description
Creates a new runner for reinforcement learning.

### Method
`new`

### Parameters
#### Arguments
- `directory` (impl AsRef<Path>) - The directory to save the checkpoints.
- `env_initializer` (EI) - Specifies how to initialize the environment.
```

--------------------------------

### AdamWStateItem Search Examples

Source: https://burn.dev/docs/burn/optim/struct.AdamWStateItem.html?search=

Examples of search queries for AdamWStateItem.

```APIDOC
## AdamWStateItem Search

### Description
Provides examples of how to search for `AdamWStateItem` within the Burn optimizer.

### Example Searches

*   `std::vec`
*   `u32 -> bool`
*   `Option<T>, (T -> U) -> Option<U>`
```

--------------------------------

### FileApplicationLoggerInstaller::install

Source: https://burn.dev/docs/burn/train/struct.FileApplicationLoggerInstaller.html

Installs the configured file application logger. This method should be called after creating an instance of FileApplicationLoggerInstaller.

```APIDOC
## fn install(&self) -> Result<(), String>

### Description
Install the application logger.

### Returns
- `Ok(())` if the logger was installed successfully.
- `Err(String)` if an error occurred during installation.
```

--------------------------------

### KLDivLossConfig::init()

Source: https://burn.dev/docs/burn/nn/loss/struct.KLDivLossConfig.html?search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

Initializes the KLDiv Loss based on the current configuration.

```APIDOC
## pub fn init(&self) -> KLDivLoss

Initialize KLDiv Loss.
```

--------------------------------

### FileApplicationLoggerInstaller Implementation

Source: https://burn.dev/docs/burn/train/trait.ApplicationLoggerInstaller.html?search=u32+-%3E+bool

Example of an implementation for the ApplicationLoggerInstaller trait, specifically for FileApplicationLoggerInstaller.

```rust
impl ApplicationLoggerInstaller for FileApplicationLoggerInstaller {
    // ... implementation details ...
}
```

--------------------------------

### NumericEntry Search Examples

Source: https://burn.dev/docs/burn/train/metric/enum.NumericEntry.html?search=

Examples of how to search for numeric entries.

```APIDOC
## NumericEntry Search Examples

This section provides examples of common search queries for numeric entries.

### Example Searches

*   `std::vec`
*   `u32 -> bool`
*   `Option<T>, (T -> U) -> Option<U>`
```