Translating once the transcription of the segment is complete.

[cevvalkoala]

I'm looking for an option to translate only once the transcription of the sentence is complete. Changing the translation multiple times makes it a bit difficult to follow.
On the same lines, an option to display transcription only once the sentence (segment) is complete would be awesome. Am I missing something?

[QuentinFuxa]

Hi, yess having a stable (and faster) translation is work in progress, will be in the next release

[cevvalkoala]

NLLW is such a useful tool. Thank you for all the hard work you put in this. I owe you a very good Turkish coffee if you ever visit Istanbul.

On a similar line, I'm trying to combine WhisperLiveKit's transcription output with DeepL's API. Stricter rules for segmentation of transcription output would make my endeavor much easier. Now the transcription output appears in huge paragraphs like the ones in the image.

Sentence by sentence presentation of the transcription output would make whisperlivekit even more awesome.

[yanrbts]

WARNING:whisperlivekit.audio_processor:Exception in results_formatter. Traceback: Traceback (most recent call last):
File "/home/yanruibing/src/WhisperLiveKit/whisperlivekit/audio_processor.py", line 470, in results_formatter
lines, undiarized_text = format_output(
^^^^^^^^^^^^^^
File "/home/yanruibing/src/WhisperLiveKit/whisperlivekit/results_formater.py", line 225, in format_output
for ts in translation_validated_segments:
TypeError: 'TimedText' object is not iterable

INFO:whisperlivekit.audio_processor:internal_buffer=0.00s | lag=0.21s |
INFO:whisperlivekit.audio_processor:internal_buffer=0.00s | lag=0.28s |
WARNING:whisperlivekit.audio_processor:Exception in results_formatter. Traceback: Traceback (most recent call last):
File "/home/yanruibing/src/WhisperLiveKit/whisperlivekit/audio_processor.py", line 470, in results_formatter
lines, undiarized_text = format_output(
^^^^^^^^^^^^^^
File "/home/yanruibing/src/WhisperLiveKit/whisperlivekit/results_formater.py", line 225, in format_output
for ts in translation_validated_segments:
TypeError: 'TimedText' object is not iterable