Collected sources and patterns will appear here. Add from search or the patterns library.
List<DatasetPath> & List<Weight> -> Batch<Data>
Merge and sample from multiple independent data paths dynamically during training using configured mixture weights.
Problem it solves
Unbalanced training datasets lead to model bias or forgetting of minority domains during fine-tuning.
Consumes
Emits
The real projects this mechanism was found in. Attribution is the point — this is how the best teams actually do it.