Expand description
This mod defines the traits for neural network modules and parameters.
This is to separate the trait definitions from the implementations, which are in the mlx-nn
crate. This also allows using the mlx_macros::ModuleParameters derive macro in crates other
than mlx-nn.
Structs§
- Param
- A simple wrapper for a module parameter.
Traits§
- Module
- Trait for a neural network module.
- Module
Parameters - Trait for accessing and updating module parameters.
- Module
Parameters Ext - Extension trait for
ModuleParameters. This is implemented for all types that implementModuleParameters. - Parameter
- Trait for a module parameter.
- Unary
Module - Marker trait for a unary neural network module.
Functions§
- update_
parameters - Update the module parameters from an iterator of (key, value) tuples.
Type Aliases§
- Flattened
Module Param - Type alias for flattened module parameters.
- Flattened
Module Param Mut - Type alias for mutably borrowed flattened module parameters.
- Flattened
Module Param Ref - Type alias for borrowed flattened module parameters.
- Module
Param - Type alias for owned module parameters.
- Module
Param Mut - Type alias for mutably borrowed module parameters.
- Module
Param Ref - Type alias for borrowed module parameters.