dasemu/Transcriptarr
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Made /asr take the encode arg correctly.

Browse Source
This commit is contained in:
McCloudS
2024-08-14 14:55:11 -06:00
committed by GitHub
parent 5f604f2ebe
commit 0fff553b0d
1 changed files with 15 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

21
subgen.py
View File

@@ -456,14 +456,23 @@ async def asr(
task_id = { 'path': f"Bazarr-asr-{random_name}" } task_id = { 'path': f"Bazarr-asr-{random_name}" }
task_queue.put(task_id) task_queue.put(task_id)
audio_data = np.frombuffer(audio_file.file.read(), np.int16).flatten().astype(np.float32) / 32768.0 args = {}
args['progress_callback'] = progress
if model_prompt: if not encode:
custom_prompt = greetings_translations.get(language, '') or custom_model_prompt args['audio'] = np.frombuffer(audio_file.file.read(), np.int16).flatten().astype(np.float32) / 32768.0
if custom_regroup: args['input_sr'] = 16000
result = model.transcribe_stable(audio_data, task=task, input_sr=16000, language=language, progress_callback=progress, initial_prompt=custom_prompt, regroup=custom_regroup, **kwargs)
else: else:
result = model.transcribe_stable(audio_data, task=task, input_sr=16000, language=language, progress_callback=progress, initial_prompt=custom_prompt, **kwargs) args['audio'] = audio_file.file.read()
if model_prompt:
args['initial_prompt'] = greetings_translations.get(language, '') or custom_model_prompt
if custom_regroup:
args['regroup'] = custom_regroup
kwargs.update(args)
result = model.transcribe_stable(task=task, language=language, **kwargs)
appendLine(result) appendLine(result)
elapsed_time = time.time() - start_time elapsed_time = time.time() - start_time
minutes, seconds = divmod(int(elapsed_time), 60) minutes, seconds = divmod(int(elapsed_time), 60)