Xabier de Zuazo a84d7ea904 Fix beam search with batch processing in Whisper decoding ...

* It ensures that audio features are correctly duplicated across beams for each batch item.
* Added a test for `decode()` that includes a regression test for this.
* Update *.github/workflows/test.yml* to run the new test for `decode()` in tiny.
* This issue was introduced in PR #1483.

2024-06-01 09:37:32 +02:00

..

conftest.py

word-level timestamps in transcribe() (#869)

2023-03-06 14:00:49 -08:00

jfk.flac

initial commit

2022-09-22 01:09:43 +09:00

test_audio.py

apply formatting with black (#1038)

2023-03-06 15:50:37 -08:00

test_decode.py

Fix beam search with batch processing in Whisper decoding

2024-06-01 09:37:32 +02:00

test_normalizer.py

apply formatting with black (#1038)

2023-03-06 15:50:37 -08:00

test_timing.py

apply formatting with black (#1038)

2023-03-06 15:50:37 -08:00

test_tokenizer.py

remove tiktoken pin (#1759)

2023-11-06 03:05:21 -08:00

test_transcribe.py

large-v3 (#1761)

2023-11-06 10:10:30 -08:00