CompressionFrontierPoint<T>

Models & Types · AiDotNet.ModelCompression

One operating point on the compression-versus-accuracy frontier: the configured strategy applied to a magnitude-ranked fraction of the model’s weights, with the resulting model re-evaluated on real data.

For Beginners

Compressing more of a model makes it smaller but usually less accurate. Each point here answers “if I compress this fraction of the weights, how small does the model get and how much accuracy do I keep?” — so you can pick the trade-off you want instead of guessing.

Constructors

Constructor Summary
CompressionFrontierPoint()

Properties

Property Summary
double AccuracyRetained { get; set; } The share of the original model’s fit retained after compressing and rebuilding, 0..1 — measured by re-running the decompressed model on the prepared data, not by a proxy.
double CompressionRatio { get; set; } Original bytes divided by compressed bytes at this point (higher means smaller).
double Fraction { get; set; } The fraction of weights (smallest-magnitude first) compressed at this operating point, 0..1.
double ReconstructionError { get; set; } Root-mean-square error between the original weights and the decompressed weights at this point.
long CompressedSizeBytes { get; set; } The compressed size in bytes at this point (compressed fraction plus the exact remainder).