Module record

Expand description

Module for the recorder.

Structs§

BinBytesRecorder: In memory recorder using the bincode format.
BinFileRecorderstd: File recorder using the bincode format.
BinGzFileRecorderstd: File recorder using the bincode format compressed with gzip.
BurnMetadata: Metadata of a record.
BurnRecord: Record that can be saved by a Recorder.
BurnRecordNoItem: Record that can be saved by a Recorder without the item.
DoublePrecisionSettings: Precision settings optimized for precision.
FullPrecisionSettings: Default precision settings.
HalfPrecisionSettings: Precision settings optimized for compactness.
JsonGzFileRecorderstd: File recorder using the json format compressed with gzip.
NamedMpkBytesRecorderstd: In memory recorder using the Named MessagePack.
NamedMpkFileRecorderstd: File recorder using the named msgpack format.
NamedMpkGzFileRecorderstd: File recorder using the named msgpack format compressed with gzip.
ParamSerde: (De)serialize parameters into a clean format.
PrettyJsonFileRecorderstd: File recorder using pretty json format for easy readability.

BytesRecorder: Recorder trait specialized to save and load data to and from bytes.
FileRecorderstd: Recorder trait specialized to save and load data to and from files.
PrecisionSettings: Settings allowing to control the precision when (de)serializing items.
Record: Trait to define a family of types which can be recorded using any settings.
Recorder: Record any item implementing Serialize and DeserializeOwned.

CompactRecorderstd: Recorder optimized for compactness.
DebugRecordSettingsstd: Debug recorder.
DefaultFileRecorderstd: Default file recorder.
DefaultRecorderstd: Default recorder.
NoStdInferenceRecorder: Inference recorder compatible with no-std.
NoStdTrainingRecorderstd: Training recorder compatible with no-std inference.
SensitiveCompactRecorderstd: Recorder optimized for compactness making it a good choice for model deployment.