Crate axonml_data 

axonml-data - Data Loading Utilities

Provides data loading infrastructure for training neural networks:

• Dataset trait for defining data sources
• DataLoader for batched iteration with parallel loading
• Samplers for controlling data access patterns
• Transforms for data augmentation

Example

use axonml_data::prelude::*;

// Define a simple dataset
struct MyDataset {
    data: Vec<(Tensor<f32>, Tensor<f32>)>,
}

impl Dataset for MyDataset {
    type Item = (Tensor<f32>, Tensor<f32>);

    fn len(&self) -> usize {
        self.data.len()
    }

    fn get(&self, index: usize) -> Option<Self::Item> {
        self.data.get(index).cloned()
    }
}

// Create a DataLoader
let loader = DataLoader::new(dataset, 32)
    .shuffle(true)
    .num_workers(4);

for batch in loader.iter() {
    // Process batch
}

@version 0.1.0 @author AutomataNexus Development Team

Re-exports

pub use collate::Collate;

pub use collate::DefaultCollate;

pub use collate::StackCollate;

pub use dataloader::Batch;

pub use dataloader::DataLoader;

pub use dataloader::DataLoaderIter;

pub use dataset::ConcatDataset;

pub use dataset::Dataset;

pub use dataset::InMemoryDataset;

pub use dataset::MapDataset;

pub use dataset::SubsetDataset;

pub use dataset::TensorDataset;

pub use sampler::BatchSampler;

pub use sampler::RandomSampler;

pub use sampler::Sampler;

pub use sampler::SequentialSampler;

pub use sampler::SubsetRandomSampler;

pub use sampler::WeightedRandomSampler;

pub use transforms::Compose;

pub use transforms::Normalize;

pub use transforms::RandomNoise;

pub use transforms::ToTensor;

pub use transforms::Transform;

Modules

collate

Collate - Batch Assembly Functions

dataloader

DataLoader - Batched Data Iteration

dataset

Dataset Trait - Core Data Abstraction

prelude

Common imports for data loading.

sampler

Samplers - Data Access Patterns

transforms

Transforms - Data Augmentation and Preprocessing