Expand description
§Burn Central Runtime
This crate is the middle layer of the Burn Central SDK. It communicate directly with the generate crate created by Burn Central CLI and provide the necessary building block to execute training and inference routines.
§Features
- Extractor for training and inference arguments.
- Wrapper for return type of training and inference routines.
- Executor to run training and inference routines.
- Error handling for runtime execution.
§Hidden element
This crate also expose some hidden elements that are only to be used by the code generated by Burn Central CLI. These elements are not meant to be used directly by the end user.
Structs§
- Args
- Args are wrapper around the config you want to inject.
- Artifact
Loader - Artifact loader for loading artifacts from Burn Central. It allow to fecth for instance other experiment endpoint to be able to restart from a certain point your experiment.
- Model
- Wrapper around the model returned by a routine.
- Multi
Device - Wrapper around multiple devices.
Enums§
Traits§
- Experiment
Args - Trait for experiments arguments. It specify that the type must be serializable, deserializable and implement default. The reason it must implement default is that when you override a value it will only override the value you provide, the rest will be filled with the default value.