github/tinygrad
1
1
Fork 0
mirror of https://github.com/tinygrad/tinygrad.git synced 2026-01-09 15:08:02 -05:00
Code Issues Packages Projects Releases Wiki Activity

whisper: support audio >30s (#2378)

Browse Source 
* whisper: support audio >30s

* make prompt indexing consistent with reference repo

* fix online
This commit is contained in:
mmmkkaaayy
2023-11-21 14:37:51 -08:00
committed by GitHub
parent 7220f5c9fc
commit 7f0cc4a4e8
3 changed files with 97 additions and 61 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
test/external/external_test_whisper_librispeech.py vendored
View File

@@ -36,7 +36,7 @@ def run_evaluation(model_name, tinygrad_expected_wer, reference_wer):
references = []
for audio, texts in tqdm.tqdm(loader):
transcriptions = transcribe_waveform(model, enc, audio.numpy())
transcriptions = transcribe_waveform(model, enc, audio.numpy(), truncate=True)
hypotheses.extend(transcriptions)
references.extend(texts)