tokenizer.py: Stating when language is not detected out of the audio

bphd · Dec 18, 2024 · 57899a3 · 57899a3
1 parent 90db0de
commit 57899a3
Showing 1 changed file with 1 addition and 1 deletion.
diff --git a/whisper/tokenizer.py b/whisper/tokenizer.py
@@ -381,7 +381,7 @@ def get_tokenizer(
 
     if multilingual:
         encoding_name = "multilingual"
-        language = language or "en"
+        language = language or "default"
         task = task or "transcribe"
     else:
         encoding_name = "gpt2"