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

update

Browse Source
This commit is contained in:
游雁 2024-03-26 15:07:45 +08:00
parent cc47491099
commit abf5af40e9
2 changed files with 14 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
funasr/__init__.py
View File

@ -31,4 +31,6 @@ def import_submodules(package, recursive=True):
import_submodules(__name__) import_submodules(__name__)
from funasr.auto.auto_model import AutoModel from funasr.auto.auto_model import AutoModel
from funasr.auto.auto_frontend import AutoFrontend from funasr.auto.auto_frontend import AutoFrontend
os.environ["HYDRA_FULL_ERROR"] = 1

15
funasr/utils/load_utils.py
View File

@ -51,13 +51,20 @@ def load_audio_text_image_video(data_or_path_or_list, fs: int = 16000, audio_fs:
if isinstance(data_or_path_or_list, str) and os.path.exists(data_or_path_or_list): # local file if isinstance(data_or_path_or_list, str) and os.path.exists(data_or_path_or_list): # local file
if data_type is None or data_type == "sound": if data_type is None or data_type == "sound":
if use_ffmpeg: # if use_ffmpeg:
data_or_path_or_list = _load_audio_ffmpeg(data_or_path_or_list, sr=fs) # data_or_path_or_list = _load_audio_ffmpeg(data_or_path_or_list, sr=fs)
data_or_path_or_list = torch.from_numpy(data_or_path_or_list).squeeze() # [n_samples,] # data_or_path_or_list = torch.from_numpy(data_or_path_or_list).squeeze() # [n_samples,]
else: # else:
# data_or_path_or_list, audio_fs = torchaudio.load(data_or_path_or_list)
# if kwargs.get("reduce_channels", True):
# data_or_path_or_list = data_or_path_or_list.mean(0)
try:
data_or_path_or_list, audio_fs = torchaudio.load(data_or_path_or_list) data_or_path_or_list, audio_fs = torchaudio.load(data_or_path_or_list)
if kwargs.get("reduce_channels", True): if kwargs.get("reduce_channels", True):
data_or_path_or_list = data_or_path_or_list.mean(0) data_or_path_or_list = data_or_path_or_list.mean(0)
except:
data_or_path_or_list = _load_audio_ffmpeg(data_or_path_or_list, sr=fs)
data_or_path_or_list = torch.from_numpy(data_or_path_or_list).squeeze() # [n_samples,]
elif data_type == "text" and tokenizer is not None: elif data_type == "text" and tokenizer is not None:
data_or_path_or_list = tokenizer.encode(data_or_path_or_list) data_or_path_or_list = tokenizer.encode(data_or_path_or_list)
elif data_type == "image": # undo elif data_type == "image": # undo