evaluateFasta: Evaluates a trained model on .fasta/fastq files

In hiddengenome/altum: deepG

Description

Returns accuracies per batch and overall confusion matrix. Evaluates batch.size * numberOfBatches samples.

Usage

evaluateFasta(
  fasta.path,
  model.path,
  batch.size = 100,
  step = 1,
  vocabulary = c("a", "c", "g", "t"),
  label_vocabulary = c("a", "c", "g", "t"),
  numberOfBatches = 10,
  filePath = NULL,
  format = "fasta",
  filename = "",
  plot = TRUE,
  mode = "lm"
)

Arguments

fasta.path	Input directory where fasta/fastq files are located.
model.path	Path to pretrained model.
batch.size	Number of samples per batch.
step	How often to take a sample.
vocabulary	Vector of allowed characters, character outside vocabulary get encoded as 0-vector.
label_vocabulary	Labels for targets. Equal to vocabulary if not given.
numberOfBatches	How many batches to evaluate.
filePath	Where to store output, if missing output won't be written.
format	File format, "fasta" or "fastq".
filename	Name of output file.
plot	Returns density plot of accuracies if TRUE.
mode	Either "lm" for language model and "label_header" or "label_folder" for label classification.
seqStart	Inserts character at beginning of sequence from one file.
seqEnd	Insert character at end of sequence from one file.
withinFile	Insert characters between fasta entries.

hiddengenome/altum documentation built on April 22, 2020, 9:33 p.m.