cjones/WhisperLive
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Minor fixes for better punctuations

Browse Source 
Signed-off-by: makaveli10 <vineet.suryan@collabora.com>
This commit is contained in:
makaveli10
2024-11-19 02:13:11 -05:00
parent e275d34943
commit a6523b6b71
1 changed files with 12 additions and 10 deletions
Download Patch File Download Diff File Expand all files Collapse all files
whisper_live/server.py
+12 -10
View File
@@ -417,7 +417,7 @@ class ServeClientBase(object):
self.prev_out = ''
self.t_start = None
self.exit = False
self.same_output_threshold = 0
self.same_output_count = 0
self.show_prev_out_thresh = 5 # if pause(no output from whisper) show previous output for 5 seconds
self.add_pause_thresh = 3 # add a blank to segment list as a pause(no speech) for 3 seconds
self.transcript = []
@@ -794,6 +794,7 @@ class ServeClientFasterWhisper(ServeClientBase):
self.initial_prompt = initial_prompt
self.vad_parameters = vad_parameters or {"onset": 0.5}
self.no_speech_thresh = 0.45
self.same_output_threshold = 10
device = "cuda" if torch.cuda.is_available() else "cpu"
if device == "cuda":
@@ -1051,7 +1052,7 @@ class ServeClientFasterWhisper(ServeClientBase):
last_segment = None
# process complete segments
if len(segments) > 1:
if len(segments) > 1 and segments[-1].no_speech_prob <= self.no_speech_thresh:
for i, s in enumerate(segments[:-1]):
text_ = s.text
self.text.append(text_)
@@ -1065,7 +1066,7 @@ class ServeClientFasterWhisper(ServeClientBase):
self.transcript.append(self.format_segment(start, end, text_, completed=True))
offset = min(duration, s.end)
# only process the segments if it satisfies the no_speech_thresh
# only process the last segment if it satisfies the no_speech_thresh
if segments[-1].no_speech_prob <= self.no_speech_thresh:
self.current_out += segments[-1].text
last_segment = self.format_segment(
@@ -1075,14 +1076,15 @@ class ServeClientFasterWhisper(ServeClientBase):
completed=False
)
if self.current_out.strip() == self.prev_out.strip() and self.current_out != '':
self.same_output_count += 1
time.sleep(0.1) # wait for some voice activity just in case there is an unitended pause from the speaker for better punctuations.
else:
self.same_output_count = 0
# if same incomplete segment is seen multiple times then update the offset
# and append the segment to the list
if self.current_out.strip() == self.prev_out.strip() and self.current_out != '':
self.same_output_threshold += 1
else:
self.same_output_threshold = 0
if self.same_output_threshold > 5:
if self.same_output_count > self.same_output_threshold:
if not len(self.text) or self.text[-1].strip().lower() != self.current_out.strip().lower():
self.text.append(self.current_out)
self.transcript.append(self.format_segment(
@@ -1093,7 +1095,7 @@ class ServeClientFasterWhisper(ServeClientBase):
))
self.current_out = ''
offset = duration
self.same_output_threshold = 0
self.same_output_count = 0
last_segment = None
else:
self.prev_out = self.current_out