Crate hpt_macros

Tensor Macros

This crate provides a set of macros to generate code for tensor operations. These macros are used to simplify and automate common tasks such as defining tensor operations, reducing dimensionality, and optimizing numerical computations.

Examples

Here’s an example of using a macro from this crate:

// Example code using a macro from this crate

Macros

conv2d_microkernel_declare_const

declare const values

conv2d_microkernel_gen_inps

generate conv2d inps

conv2d_microkernel_gen_kernels

generate conv2d kernels

conv2d_microkernel_gen_pad_inps

generate conv2d inps

conv2d_microkernel_gen_results

generate conv2d repeat results

dwconv2d_microkernel_gen_pad_inps

generate conv2d inps

dwconv2d_microkernel_gen_results

generate conv2d repeat results

float_out_binary

implement float out binary trait

float_out_binary_simd

implement simd float out binary trait

float_out_binary_simd_with_lhs_scalar

implement simd float out binary trait with lhs scalar

float_out_binary_simd_with_rhs_scalar

implement simd float out binary trait with rhs scalar

float_out_unary

implement float out unary trait

gen_fast_layernorm_simd_helper

generate fast layernorm simd helper

gen_fast_reduce_simd_helper

generate fast reduce simd helper

gen_reduce_dim_not_include_simd_helper

generate reduce dim not include simd helper

impl_bitwise_out

implement bitwise out trait

impl_cmp

implement compare trait

impl_cmp_cuda

implement compare trait

impl_cuda_bitwise_out

implement bitwise out trait

impl_eval

implement eval trait

impl_from_scalar

implement from scalar trait

impl_into_scalar

implment into scalar trait

impl_into_vec

implment into vec trait

impl_normal_out_binary

generate notmal out trait

impl_normal_out_simd

implement simd normal out trait

impl_normal_out_simd_with_lhs_scalar

implement simd normal out trait with lhs scalar

impl_normal_out_simd_with_rhs_scalar

implement simd normal out trait with rhs scalar

impl_normal_out_unary

gemerate normal out unary trait

impl_normal_out_unary_simd

gemerate normal out unary trait

impl_scalar_convert

implement scalar convert trait

maxpool2d_microkernel_gen_results

generate maxpool2d kernels generate conv2d repeat results

pwconv2d_microkernel_gen_pad_inps

generate pwconv2d inps

select

parse the input and generate the corresponding slice

simd_bitwise

implement simd bitwise trait

simd_cmp

implement simd cmp trait

simd_eval

implement simd eval trait

simd_float_out_unary

implement simd float out unary trait

Derive Macros

FromSafeTensors

generate from safetensors trait

Load

generate load trait

Save

generate save trait