@@ -123,7 +123,7 @@ def __init__(
123123 interim_results : bool = True ,
124124 punctuate : bool = True ,
125125 spoken_punctuation : bool = False ,
126- enable_word_time_offsets : bool = True ,
126+ enable_word_time_offsets : NotGivenOr [ bool ] = NOT_GIVEN ,
127127 enable_word_confidence : bool = False ,
128128 enable_voice_activity_events : bool = False ,
129129 model : SpeechModels | str = "latest_long" ,
@@ -148,7 +148,7 @@ def __init__(
148148 interim_results(bool): whether to return interim results (default: True)
149149 punctuate(bool): whether to punctuate the audio (default: True)
150150 spoken_punctuation(bool): whether to use spoken punctuation (default: False)
151- enable_word_time_offsets(bool): whether to enable word time offsets (default: True )
151+ enable_word_time_offsets(bool): whether to enable word time offsets (default: None )
152152 enable_word_confidence(bool): whether to enable word confidence (default: False)
153153 enable_voice_activity_events(bool): whether to enable voice activity events (default: False)
154154 model(SpeechModels): the model to use for recognition default: "latest_long"
@@ -163,6 +163,18 @@ def __init__(
163163 """
164164 if not is_given (use_streaming ):
165165 use_streaming = True
166+
167+ if model == "chirp_3" :
168+ if is_given (enable_word_time_offsets ) and enable_word_time_offsets :
169+ logger .warning (
170+ "Chirp 3 does not support word timestamps, setting 'enable_word_time_offsets' to False."
171+ )
172+ enable_word_time_offsets = False
173+ elif is_given (enable_word_time_offsets ):
174+ enable_word_time_offsets = enable_word_time_offsets
175+ else :
176+ enable_word_time_offsets = True
177+
166178 super ().__init__ (
167179 capabilities = stt .STTCapabilities (
168180 streaming = use_streaming ,
0 commit comments