sotastream package
Subpackages
- sotastream.augmentors package
- sotastream.filters package
- sotastream.pipelines package
- sotastream.utils package
Submodules
Module contents
- class sotastream.Defaults[source]
Bases:
objectDefault values for pipeline arguments
- BUFFER_SIZE = 1000000
- DOC_PROB = 0.0
- DOC_PROB_PARALLEL = 0.0
- DOC_SEPARATOR = ' <eos>'
- MAX_TOKENS = 250
- NUM_PROCESSES = 16
- QUEUE_BUFFER_SIZE = 10000
- QUIET = False
- SAMPLE_FILE = None
- SAMPLE_LENGTH = True
- SEED = 0
- SEPARATOR = ' '
- SHUFFLE = True
- SPM_MODEL = None