KGEmbeddingTrainingResult
Models & Types · AiDotNet.RetrievalAugmentedGeneration.Graph.Embeddings
Contains the results of training a knowledge graph embedding model.
For Beginners
After training, this tells you:
- EpochLosses: How the error decreased over time (should go down)
- EntityCount/RelationCount: How many entities and relation types were learned
- TripleCount: Total number of facts used for training
- TrainingDuration: How long training took
How It Works
This class captures training metrics including per-epoch loss values, entity/relation counts, and training duration for diagnostics and evaluation.
Constructors
| Constructor | Summary |
|---|---|
KGEmbeddingTrainingResult() |
Properties
| Property | Summary |
|---|---|
List<double> EpochLosses { get; set; } |
Average loss value for each training epoch. |
TimeSpan TrainingDuration { get; set; } |
Wall-clock training duration. |
int EntityCount { get; set; } |
Number of unique entities in the training data. |
int RelationCount { get; set; } |
Number of unique relation types in the training data. |
int TotalEpochs { get; set; } |
Total number of epochs completed. |
int TripleCount { get; set; } |
Total number of triples used for training. |