Merge pull request #259 from aidancrowther/fix-language-detection

Correctly load n_dims for openai_whisper model

CHANGELOG.md

@@ -4,6 +4,10 @@ Changelog
 Unreleased
 ----------
 
+### Added
+
+  - Set mel generation to adjust n_dims automatically to match the loaded model
+
 [1.6.0] (2024-10-06)
 --------------------
 

app/openai_whisper/core.py

@@ -48,7 +48,7 @@ def language_detection(audio):
     audio = whisper.pad_or_trim(audio)
 
     # make log-Mel spectrogram and move to the same device as the model
-    mel = whisper.log_mel_spectrogram(audio).to(model.device)
+    mel = whisper.log_mel_spectrogram(audio, model.dims.n_mels).to(model.device)
 
     # detect the spoken language
     with model_lock: