chiiyeh
Committed by GitHub

add hotwords docstring to offline_recognizer and online_recognizer (#546)

@@ -82,6 +82,12 @@ class OfflineRecognizer(object): @@ -82,6 +82,12 @@ class OfflineRecognizer(object):
82 max_active_paths: 82 max_active_paths:
83 Maximum number of active paths to keep. Used only when 83 Maximum number of active paths to keep. Used only when
84 decoding_method is modified_beam_search. 84 decoding_method is modified_beam_search.
  85 + hotwords_file:
  86 + The file containing hotwords, one words/phrases per line, and for each
  87 + phrase the bpe/cjkchar are separated by a space.
  88 + hotwords_score:
  89 + The hotword score of each token for biasing word/phrase. Used only if
  90 + hotwords_file is given with modified_beam_search as decoding method.
85 blank_penalty: 91 blank_penalty:
86 The penalty applied on blank symbol during decoding. 92 The penalty applied on blank symbol during decoding.
87 debug: 93 debug:
@@ -100,6 +100,12 @@ class OnlineRecognizer(object): @@ -100,6 +100,12 @@ class OnlineRecognizer(object):
100 max_active_paths: 100 max_active_paths:
101 Use only when decoding_method is modified_beam_search. It specifies 101 Use only when decoding_method is modified_beam_search. It specifies
102 the maximum number of active paths during beam search. 102 the maximum number of active paths during beam search.
  103 + hotwords_file:
  104 + The file containing hotwords, one words/phrases per line, and for each
  105 + phrase the bpe/cjkchar are separated by a space.
  106 + hotwords_score:
  107 + The hotword score of each token for biasing word/phrase. Used only if
  108 + hotwords_file is given with modified_beam_search as decoding method.
103 provider: 109 provider:
104 onnxruntime execution providers. Valid values are: cpu, cuda, coreml. 110 onnxruntime execution providers. Valid values are: cpu, cuda, coreml.
105 model_type: 111 model_type: