kotoba-tech
/

kotoba-whisper-v1.0

Automatic Speech Recognition

hf-asr-leaderboard

Model card Files Files and versions

asahi417 commited on May 8, 2024

Commit

a345421

·

verified ·

1 Parent(s): fab38e1

Update benchmark.py

Files changed (1) hide show

benchmark.py +1 -1

benchmark.py CHANGED Viewed

@@ -12,7 +12,7 @@ model_id = "kotoba-tech/kotoba-whisper-v1.0"
 pipe = pipeline(
     "automatic-speech-recognition",
     model=model_id,
-    torch_dtype=torch.bfloat16
 )
 test_audio = [

 pipe = pipeline(
     "automatic-speech-recognition",
     model=model_id,
+    torch_dtype=torch.float32
 )
 test_audio = [