A high-performance inference engine for AI models
use crate::backends::{
common::{Kernels, kernel::Unsupported},
cpu::Cpu,
};
mod activation;
pub(crate) mod activation_transform;
mod attention;
mod confidence_head;
mod convolution;
mod embedding;
mod gated_act_mul;
mod gdn;
mod logit_transform;
mod matmul;
mod moe;
mod normalization;
mod pooling;
mod radix_top_k_small;
mod sampling;
mod short_conv;
mod tensor_add_bias;
mod tensor_add_scale;
mod weaver;
include!(concat!(env!("OUT_DIR"), "/cpu.rs"));
pub struct CpuKernels;
impl Kernels for CpuKernels {
type Backend = Cpu;
autogen_kernels!();
type AttentionKernel = attention::AttentionCpuKernel;
type DeltaNetChunkedPrefill = Unsupported<Cpu>;
type DeltaNetTreeVerify = Unsupported<Cpu>;
type MatmulKernel = matmul::MatmulCpuKernel;
type RadixTopKSmall = radix_top_k_small::CpuRadixTopKSmall;
}