MCPcopy Create free account
hub / github.com/catboost/catboost / quantize

Function quantize

catboost/python-package/catboost/utils.py:541–727  ·  view source on GitHub ↗

Construct quantized Pool from non-quantized pool stored in file. This method does not load whole non-quantized source dataset into memory so it can be used for huge datasets that fit in memory only after quantization. Parameters ---------- data_path : string or pathlib.Path

(
    data_path,
    column_description=None,
    pairs=None,
    graph=None,
    delimiter='\t',
    has_header=False,
    ignore_csv_quoting=False,
    feature_names=None,
    thread_count=-1,
    ignored_features=None,
    per_float_feature_quantization=None,
    border_count=None,
    max_bin=None,
    feature_border_type=None,
    nan_mode=None,
    input_borders=None,
    task_type=None,
    used_ram_limit=None,
    random_seed=None,
    log_cout=sys.stdout,
    log_cerr=sys.stderr,
    **kwargs
)

Source from the content-addressed store, hash-verified

539
540
541def quantize(
542 data_path,
543 column_description=None,
544 pairs=None,
545 graph=None,
546 delimiter='\t',
547 has_header=False,
548 ignore_csv_quoting=False,
549 feature_names=None,
550 thread_count=-1,
551 ignored_features=None,
552 per_float_feature_quantization=None,
553 border_count=None,
554 max_bin=None,
555 feature_border_type=None,
556 nan_mode=None,
557 input_borders=None,
558 task_type=None,
559 used_ram_limit=None,
560 random_seed=None,
561 log_cout=sys.stdout,
562 log_cerr=sys.stderr,
563 **kwargs
564):
565 """
566 Construct quantized Pool from non-quantized pool stored in file.
567 This method does not load whole non-quantized source dataset into memory
568 so it can be used for huge datasets that fit in memory only after quantization.
569
570 Parameters
571 ----------
572 data_path : string or pathlib.Path
573 Path (with optional scheme) to non-quantized dataset.
574
575 column_description : string, [default=None]
576 ColumnsDescription parameter.
577 There are several columns description types: Label, Categ, Num, Auxiliary, DocId, Weight, Baseline, GroupId, Timestamp.
578 All columns are Num as default, it's not necessary to specify
579 this type of columns. Default Label column index is 0 (zero).
580 If None, Label column is 0 (zero) as default, all data columns are Num as default.
581 If string or pathlib.Path, giving the path to the file with ColumnsDescription in column_description format.
582
583 pairs : string or pathlib.Path, [default=None]
584 Path to the file with pairs description.
585
586 graph : string or pathlib.Path, [default=None]
587 Path to the file with graph description.
588
589 has_header : bool, [default=False]
590 If True, read column names from first line.
591
592 ignore_csv_quoting : bool optional (default=False)
593 If True ignore quoting '"'.
594
595 feature_names : string or pathlib.Path, [default=None]
596 Path with scheme for feature names data to load.
597
598 thread_count : int, [default=-1]

Callers 3

QuantizeFunction · 0.50

Calls 8

_readMethod · 0.95
isinstanceFunction · 0.85
_process_synonymsFunction · 0.85
PoolClass · 0.70
CatBoostErrorClass · 0.50
popMethod · 0.45
formatMethod · 0.45