chrissewell/whisperX
1
0
Fork 0
mirror of https://github.com/m-bain/whisperX.git synced 2025-07-01 18:17:27 -04:00
Code Issues Packages Projects Releases Wiki Activity

Merge pull request #96 from smly/fix-batch-processing

Browse Source 
FIX: Assertion error in batch processing
This commit is contained in:
m-bain
2023-02-22 12:11:01 +00:00
committed by GitHub
parent 2b1ffa12b8 27fe502344
commit 847a3cd85b
1 changed files with 21 additions and 26 deletions
Download Patch File Download Diff File Expand all files Collapse all files

47
whisperx/transcribe.py
View File

@ -365,8 +365,6 @@ def transcribe_with_vad_parallel(
if mel is None: if mel is None:
mel = log_mel_spectrogram(audio) mel = log_mel_spectrogram(audio)
output = {"segments": []}
vad_segments = vad_pipeline(audio) vad_segments = vad_pipeline(audio)
# merge segments to approx 30s inputs to make whisper most appropraite # merge segments to approx 30s inputs to make whisper most appropraite
vad_segments = merge_chunks(vad_segments) vad_segments = merge_chunks(vad_segments)
@ -428,7 +426,9 @@ def transcribe_with_vad_parallel(
language = kwargs["language"] language = kwargs["language"]
task = kwargs["task"] task = kwargs["task"]
tokenizer = get_tokenizer(model.is_multilingual, language=language, task=task) tokenizer = get_tokenizer(model.is_multilingual, language=language, task=task)
result_segments = post_process_results(
output = post_process_results(
vad_segments,
decode_result, decode_result,
duration_list, duration_list,
offset_list, offset_list,
@ -438,29 +438,11 @@ def transcribe_with_vad_parallel(
no_speech_threshold=no_speech_threshold, no_speech_threshold=no_speech_threshold,
logprob_threshold=logprob_threshold, logprob_threshold=logprob_threshold,
verbose=verbose) verbose=verbose)
# post processing: collect outputs
assert len(result_segments) == len(vad_segments)
for sdx, (seg_t, result) in enumerate(zip(vad_segments, result_segments)):
seg_t["text"] = result["text"]
output["segments"].append(
{
"start": seg_t["start"],
"end": seg_t["end"],
"language": result["language"],
"text": result["text"],
"seg-text": [x["text"] for x in result["segments"]],
"seg-start": [x["start"] for x in result["segments"]],
"seg-end": [x["end"] for x in result["segments"]],
}
)
output["language"] = output["segments"][0]["language"]
return output return output
def post_process_results( def post_process_results(
vad_segments,
result_list, result_list,
duration_list, duration_list,
offset_list, offset_list,
@ -478,7 +460,7 @@ def post_process_results(
) # time per output token: 0.02 (seconds) ) # time per output token: 0.02 (seconds)
all_tokens = [] all_tokens = []
all_segments = [] all_segments = []
outputs = [] output = {"segments": []}
def add_segment( def add_segment(
*, start: float, end: float, text_tokens: torch.Tensor, result: DecodingResult *, start: float, end: float, text_tokens: torch.Tensor, result: DecodingResult
@ -505,7 +487,7 @@ def post_process_results(
print(f"[{format_timestamp(start)} --> {format_timestamp(end)}] {text}") print(f"[{format_timestamp(start)} --> {format_timestamp(end)}] {text}")
# process the output # process the output
for result, segment_duration, timestamp_offset in zip(result_list, duration_list, offset_list): for seg_t, result, segment_duration, timestamp_offset in zip(vad_segments, result_list, duration_list, offset_list):
all_tokens = [] all_tokens = []
all_segments = [] all_segments = []
@ -568,9 +550,22 @@ def post_process_results(
seek += segment_shape seek += segment_shape
all_tokens.extend(tokens.tolist()) all_tokens.extend(tokens.tolist())
outputs.append(dict(text=tokenizer.decode(all_tokens), segments=all_segments, language=language)) result = dict(text=tokenizer.decode(all_tokens), segments=all_segments, language=language)
output["segments"].append(
{
"start": seg_t["start"],
"end": seg_t["end"],
"language": result["language"],
"text": result["text"],
"seg-text": [x["text"] for x in result["segments"]],
"seg-start": [x["start"] for x in result["segments"]],
"seg-end": [x["end"] for x in result["segments"]],
}
)
return outputs output["language"] = output["segments"][0]["language"]
return output
def cli(): def cli():