From 24926c98e0891d2972a8b438bd0898b472d5f128 Mon Sep 17 00:00:00 2001 From: koiking213 Date: Tue, 20 Feb 2024 22:46:04 +0900 Subject: [PATCH] specify audio dtype --- whisper_online.py | 2 +- whisper_online_server.py | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/whisper_online.py b/whisper_online.py index 2794506..bcaeec4 100644 --- a/whisper_online.py +++ b/whisper_online.py @@ -11,7 +11,7 @@ import math @lru_cache def load_audio(fname): a, _ = librosa.load(fname, sr=16000) - return a + return a.astype('float32') def load_audio_chunk(fname, beg, end): audio = load_audio(fname) diff --git a/whisper_online_server.py b/whisper_online_server.py index 13a85de..a609ef0 100644 --- a/whisper_online_server.py +++ b/whisper_online_server.py @@ -138,7 +138,7 @@ class ServerProcessor: break sf = soundfile.SoundFile(io.BytesIO(raw_bytes), channels=1,endian="LITTLE",samplerate=SAMPLING_RATE, subtype="PCM_16",format="RAW") audio, _ = librosa.load(sf,sr=SAMPLING_RATE) - out.append(audio) + out.append(audio.astype('float32')) if not out: return None return np.concatenate(out)