Module batch 

Batch matrix operations for AI/ML workloads

This module provides matrix operations optimized for processing batches of data, which is especially useful for machine learning applications such as mini-batch gradient descent, convolutional neural networks, and transformer models.

Re-exports

pub use attention::batch_flash_attention;

pub use attention::batch_multi_head_attention;

pub use attention::batch_multi_query_attention;

Modules

attention

Batched attention mechanisms for transformer models

Functions

batch_add

Add a vector to each matrix in a batch

batch_matmul

Perform matrix multiplication on a batch of matrices

batch_matvec

Batch matrix-vector multiplication

batch_sum

Compute the sum of a batch of matrices