deepbinner classify

usage: deepbinner classify [--native] [--rapid] [-s START_MODEL] [-e END_MODEL]
                           [--scan_size SCAN_SIZE] [--score_diff SCORE_DIFF]
                           [--require_either] [--require_start] [--require_both]
                           [--batch_size BATCH_SIZE]
                           [--intra_op_parallelism_threads INTRA_OP_PARALLELISM_THREADS]
                           [--inter_op_parallelism_threads INTER_OP_PARALLELISM_THREADS]
                           [--device_count DEVICE_COUNT]
                           [--omp_num_threads OMP_NUM_THREADS] [--verbose] [-h]
                           input

Classify fast5 reads

Positional:
  input                     One of the following: a single fast5 file, a directory
                            of fast5 files (will be searched recursively) or a
                            tab-delimited file of training data

Model presets:
  --native                  Preset for EXP-NBD103 read start and end models
  --rapid                   Preset for SQK-RBK004 read start model

Models (at least one is required if not using a preset):
  -s START_MODEL, --start_model START_MODEL
                            Model trained on the starts of reads
  -e END_MODEL, --end_model END_MODEL
                            Model trained on the ends of reads

Barcoding:
  --scan_size SCAN_SIZE     This much of a read's start/end signal will examined
                            for barcode signals (default: 6144)
  --score_diff SCORE_DIFF   For a read to be classified, there must be this much
                            difference between the best and second-best barcode
                            scores (default: 0.5)

Two model (read start and read end) behaviour:
  --require_either          Most lenient approach: a barcode call on either the
                            start or end is sufficient to classify a read, as long
                            as they do not disagree on the barcode
  --require_start           Moderate approach: a start barcode is required to
                            classify a read but an end barcode is optional
                            (default behaviour)
  --require_both            Most stringent approach: both start and end barcodes
                            must be present and agree to classify a read

Performance:
  --batch_size BATCH_SIZE   Neural network batch size (default: 256)
  --intra_op_parallelism_threads INTRA_OP_PARALLELISM_THREADS
                            TensorFlow's intra_op_parallelism_threads config
                            option (default: 12)
  --inter_op_parallelism_threads INTER_OP_PARALLELISM_THREADS
                            TensorFlow's inter_op_parallelism_threads config
                            option (default: 1)
  --device_count DEVICE_COUNT
                            TensorFlow's device_count config option (default: 1)
  --omp_num_threads OMP_NUM_THREADS
                            OMP_NUM_THREADS environment variable value (default:
                            12)

Other:
  --verbose                 Include the output probabilities for all barcodes in
                            the results (default: just show the final barcode
                            call)
  -h, --help                Show this help message and exit