speech_recognition/SenseVoice
1
0
Fork 0
mirror of https://github.com/FunAudioLLM/SenseVoice.git synced 2025-09-15 15:08:35 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity

docs

Browse Source
This commit is contained in:
游雁 2024-07-23 18:41:01 +08:00
parent dfe6e6a7bd
commit 751333fb28
4 changed files with 9 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
README.md
View File

@ -170,12 +170,13 @@ from funasr.utils.postprocess_utils import rich_transcription_postprocess
model_dir = "iic/SenseVoiceSmall" model_dir = "iic/SenseVoiceSmall"
m, kwargs = SenseVoiceSmall.from_pretrained(model=model_dir, device="cuda:0") m, kwargs = SenseVoiceSmall.from_pretrained(model=model_dir, device="cuda:0")
m.eval()
res = m.inference( res = m.inference(
data_in=f"{kwargs['model_path']}/example/en.mp3", data_in=f"{kwargs['model_path']}/example/en.mp3",
language="auto", # "zh", "en", "yue", "ja", "ko", "nospeech" language="auto", # "zh", "en", "yue", "ja", "ko", "nospeech"
use_itn=False, use_itn=False,
ban_emo_unk=False,
**kwargs, **kwargs,
) )

3
README_ja.md
View File

@ -171,12 +171,13 @@ from funasr.utils.postprocess_utils import rich_transcription_postprocess
model_dir = "iic/SenseVoiceSmall" model_dir = "iic/SenseVoiceSmall"
m, kwargs = SenseVoiceSmall.from_pretrained(model=model_dir, device="cuda:0") m, kwargs = SenseVoiceSmall.from_pretrained(model=model_dir, device="cuda:0")
m.eval()
res = m.inference( res = m.inference(
data_in=f"{kwargs['model_path']}/example/en.mp3", data_in=f"{kwargs['model_path']}/example/en.mp3",
language="auto", # "zh", "en", "yue", "ja", "ko", "nospeech" language="auto", # "zh", "en", "yue", "ja", "ko", "nospeech"
use_itn=False, use_itn=False,
ban_emo_unk=False,
**kwargs, **kwargs,
) )

5
README_zh.md
View File

@ -145,7 +145,7 @@ print(text)
- `use_itn`：输出结果中是否包含标点与逆文本正则化。 - `use_itn`：输出结果中是否包含标点与逆文本正则化。
- `batch_size_s` 表示采用动态batch，batch中总音频时长，单位为秒s。 - `batch_size_s` 表示采用动态batch，batch中总音频时长，单位为秒s。
- `merge_vad`：是否将 vad 模型切割的短音频碎片合成，合并后长度为`merge_length_s`，单位为秒s。 - `merge_vad`：是否将 vad 模型切割的短音频碎片合成，合并后长度为`merge_length_s`，单位为秒s。
- `ban_emo_unk`：禁用emo_unk标签，禁用后所有的句子都会被赋与情感标签。 - `ban_emo_unk`：禁用emo_unk标签，禁用后所有的句子都会被赋与情感标签。默认`False`
</details> </details>
如果输入均为短音频（小于30s），并且需要批量化推理，为了加快推理效率，可以移除vad模型，并设置`batch_size` 如果输入均为短音频（小于30s），并且需要批量化推理，为了加快推理效率，可以移除vad模型，并设置`batch_size`
@ -174,12 +174,13 @@ from funasr.utils.postprocess_utils import rich_transcription_postprocess
model_dir = "iic/SenseVoiceSmall" model_dir = "iic/SenseVoiceSmall"
m, kwargs = SenseVoiceSmall.from_pretrained(model=model_dir, device="cuda:0") m, kwargs = SenseVoiceSmall.from_pretrained(model=model_dir, device="cuda:0")
m.eval()
res = m.inference( res = m.inference(
data_in=f"{kwargs['model_path']}/example/en.mp3", data_in=f"{kwargs['model_path']}/example/en.mp3",
language="auto", # "zh", "en", "yue", "ja", "ko", "nospeech" language="auto", # "zh", "en", "yue", "ja", "ko", "nospeech"
use_itn=False, use_itn=False,
ban_emo_unk=False,
**kwargs, **kwargs,
) )

3
demo2.py
View File

@ -9,12 +9,13 @@ from funasr.utils.postprocess_utils import rich_transcription_postprocess
model_dir = "iic/SenseVoiceSmall" model_dir = "iic/SenseVoiceSmall"
m, kwargs = SenseVoiceSmall.from_pretrained(model=model_dir, device="cuda:0") m, kwargs = SenseVoiceSmall.from_pretrained(model=model_dir, device="cuda:0")
m.eval()
res = m.inference( res = m.inference(
data_in=f"{kwargs['model_path']}/example/en.mp3", data_in=f"{kwargs['model_path']}/example/en.mp3",
language="auto", # "zh", "en", "yue", "ja", "ko", "nospeech" language="auto", # "zh", "en", "yue", "ja", "ko", "nospeech"
use_itn=False, use_itn=False,
ban_emo_unk=False,
**kwargs, **kwargs,
) )