admin-valentin/WhisperLiveKit
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

simulstreaming: fixes token handling during warm-up phase

Browse source
This commit is contained in:
Quentin Fuxa 2025-07-31 16:25:34 +02:00
parent f85329e112
commit 8a5e2adb1e
3 changed files with 5 additions and 6 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

7
whisperlivekit/simul_whisper/simul_whisper.py
View file

@ -261,9 +261,10 @@ class PaddedAlignAttWhisper:
segments_len -= removed_len segments_len -= removed_len
self.last_attend_frame -= int(TOKENS_PER_SECOND*removed_len) self.last_attend_frame -= int(TOKENS_PER_SECOND*removed_len)
self.segments = self.segments[1:] self.segments = self.segments[1:]
logger.debug(f"remove segments: {len(self.segments)} {len(self.tokens)}") if len(self.tokens) > 1: # When warming up, we can have a too long segments_len while not having any tokens yet
self.context.append_token_ids(self.tokens[1][0,:]) self.context.append_token_ids(self.tokens[1][0,:])
self.tokens = [self.initial_tokens] + self.tokens[2:] self.context.append_token_ids(self.tokens[1][0,:])
self.tokens = [self.initial_tokens] + self.tokens[2:]
return removed_len return removed_len

1
whisperlivekit/whisper_streaming_custom/backends.py
View file

@ -481,7 +481,6 @@ class SimulStreamingASR(ASRBase):
try: try:
if isinstance(audio, np.ndarray): if isinstance(audio, np.ndarray):
audio = torch.from_numpy(audio).float() audio = torch.from_numpy(audio).float()
print(audio)
self.model.insert_audio(audio) self.model.insert_audio(audio)
self.model.infer(True) self.model.infer(True)
self.model.refresh_segment(complete=True) self.model.refresh_segment(complete=True)

3
whisperlivekit/whisper_streaming_custom/online_asr.py
View file

@ -680,8 +680,7 @@ class SimulStreamingOnlineProcessor:
except Exception as e: except Exception as e:
logger.error(f"SimulStreaming processing error: {e}") logger.exception(f"SimulStreaming processing error: {e}")
logger.error(f"Error details: {type(e).__name__}: {str(e)}")
return [], self.end return [], self.end
def finish(self) -> Tuple[List[ASRToken], float]: def finish(self) -> Tuple[List[ASRToken], float]: