Crate instmodel_inference 

Neural inference library for executing optimized computation sequences.

This library provides functionality to run neural network inference using a sequence of instructions that operate on computation buffers. It supports various operations like dot products, activations, copying, and element-wise operations.

Re-exports

pub use activation::Activation;

pub use instruction_model::InstructionModel;

pub use instruction_model_info::InstructionModelInfo;

pub use instruction_model_info::ValidationData;

pub use parallel_predict::ParallelPredictOutput;

pub use parallel_predict::PredictConfig;

Modules

activation

Activation functions for neural network operations.

benchmarks

Benchmark suite for neural inference performance testing.

errors

Error types for the neural inference library.

gpu

GPU neural network inference module.

high_performance_execution_utils

Execution graph structures for parallel instruction execution.

instruction_model

Core instruction model for neural network inference.

instruction_model_info

Data structures for instruction model configuration.

instructions

Instruction implementations for neural network operations.

parallel_predict

utils

Utility functions for the neural inference library.