ai.minxiao.ds4s.core.dl4j.learning
ESLearner
Related Doc:
package learning
model saver type, options: in-memory-model, local-file-graph, local-file-model, default="in-memory-model"
model save directory for local-file modes, default="./"
model encoding for local-file modes, default=Charset.defaultCharset options:
US-ASCII Seven-bit ASCII, a.k.a. ISO646-US, a.k.a. the Basic Latin block of the Unicode character set ISO-8859-1 ISO Latin Alphabet No. 1, a.k.a. ISO-LATIN-1 UTF-8 Eight-bit UCS Transformation Format UTF-16BE Sixteen-bit UCS Transformation Format, big-endian byte order UTF-16LE Sixteen-bit UCS Transformation Format, little-endian byte order UTF-16 Sixteen-bit UCS Transformation Format, byte order identified by an optional byte-order mark
saveLastModel whether also save the last model in additional to the best model, default=false
epoch termination conditions, default=Array("max-epochs") all possible options: "max-epochs", "best-score-epoch", "score-improvement-epoch"
max epochs for max-epochs, default=10
whether lesser is better for the score of "best-score-epoch", default=true
max epochs (consecutive epochs) without improvements, default=5
min improvement, default=1E-3
evaluate every N epochs, default=1
iteration termination conditions, default=Array("invalid-score-iteration") all possible options: "invalid-score-iteration", "max-score-iteration", "max-time-iteration"
max score for "max-score-iteration", default=1E-3
max time amount, default=30L,
score calculator, default="dataset-loss" all options: "dataset-loss", "roc-score"
dataset for score calculation, required, must be provided
whether to average the score, default=true
ROCType, default=ROCType.ROC, all options {ROC, BINARY, MULTICLASS}
ROC Metric, default=Metric.AUC, all options {AUC, AUPRC}
ESLearner