speech_recognition/FunASR
1
0
Fork 0
mirror of https://github.com/modelscope/FunASR synced 2025-09-15 14:48:36 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity

add support mixture of kaldi_ark or sound

Browse Source
This commit is contained in:
志浩 2024-09-24 19:36:18 +08:00
parent d104000f82
commit 0a65aaf266
1 changed files with 5 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
funasr/utils/load_utils.py
View File

@ -129,7 +129,10 @@ def load_audio_text_image_video(
if ".ark:" in data_or_path_or_list:
data_mat = kaldiio.load_mat(data_or_path_or_list)
if isinstance(data_mat, tuple):
audio_fs, mat = data_mat
if isinstance(data_mat[0], int):
audio_fs, mat = data_mat
else:
mat, audio_fs = data_mat
else:
mat = data_mat
if mat.dtype == "int16" or mat.dtype == "int32":
@ -137,7 +140,7 @@ def load_audio_text_image_video(
mat = mat / 32768
if mat.ndim == 2:
mat = mat[:, 0]
data_or_path_or_list = mat
data_or_path_or_list = torch.from_numpy(mat)
else:
try:
data_or_path_or_list, audio_fs = torchaudio.load(data_or_path_or_list)