oneliner-runtime 0.2.0

Runtime support for Oneliner model inference
Documentation
#![no_std]

#[cfg(feature = "alloc")]
extern crate alloc;

mod arena;
mod buffer;
mod executor;
mod interface;

#[cfg(feature = "iree-runtime")]
mod iree;

pub use buffer::{
    concurrent, fill, Access, AnyBuffer, AnyBufferRange, Buffer, BufferMut, BufferRange,
    BufferSource,
};

pub use executor::{DefaultExecutor, Executor, SequentialExecutor, WorkItem};
pub use interface::{
    Error, ModelArtifacts, ModelInference, ModelSource, Shape, Tensor, Tensor4D, TensorArray,
};
#[cfg(feature = "iree-runtime")]
pub use iree::{
    dispatch, dispatch_fn_from_library, iree_hal_executable_dispatch_state_v0_t,
    iree_hal_executable_environment_v0_t, iree_hal_executable_import_thunk_v0_t,
    iree_hal_executable_import_v0_t, iree_hal_executable_library_header_t,
    iree_hal_executable_library_query_fn_t, iree_hal_executable_workgroup_state_v0_t,
    iree_hal_processor_v0_t, try_dispatch, try_dispatch_with_executor, DispatchFn,
};

pub use aligned::{Aligned, A16, A2, A32, A4, A64};
pub type AlignedType = A64;

pub use arena::{ArenaStorage, OwnedArena, SharedArena};