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

whisper bugfix

Browse Source
This commit is contained in:
游雁 2024-08-02 01:20:34 +08:00
parent d238a5ab44
commit 7f6099140a
1 changed files with 4 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
funasr/models/whisper/model.py
View File

@ -9,6 +9,7 @@ from torch import Tensor
from torch import nn from torch import nn
import whisper import whisper
# import whisper_timestamped as whisper # import whisper_timestamped as whisper
from funasr.utils.load_utils import load_audio_text_image_video, extract_fbank from funasr.utils.load_utils import load_audio_text_image_video, extract_fbank
@ -111,10 +112,10 @@ class WhisperWarp(nn.Module):
# decode the audio # decode the audio
options = whisper.DecodingOptions(**kwargs.get("DecodingOptions", {})) options = whisper.DecodingOptions(**kwargs.get("DecodingOptions", {}))
result = whisper.decode(self.model, speech, language='english') result = whisper.decode(self.model, speech, options=options)
# result = whisper.transcribe(self.model, speech) # result = whisper.transcribe(self.model, speech)
results = [] results = []
result_i = {"key": key[0], "text": result.text} result_i = {"key": key[0], "text": result.text}