AI/MiniCPM-o-2_6
1
0
Fork 0
mirror of https://www.modelscope.cn/OpenBMB/MiniCPM-o-2_6.git synced 2025-07-31 01:05:59 +08:00
Code Issues Packages Projects Releases Wiki Activity

Release interface in processing

Browse Source
This commit is contained in:
Hongji Zhu 2025-02-07 15:00:29 +08:00
parent 64edf3d723
commit 4f835638ec
1 changed files with 26 additions and 26 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

22
processing_minicpmo.py
View File

@ -102,16 +102,7 @@ class MiniCPMOProcessor(ProcessorMixin):
return MiniCPMOBatchFeature(data={**model_inputs}) return MiniCPMOBatchFeature(data={**model_inputs})
def audio_feature_extract( def get_audio_placeholder(self, audio_lens, chunk_input, chunk_length):
self,
audios: Union[np.ndarray, List[np.ndarray], List[List[np.ndarray]]],
audio_parts: Optional[list] = None,
chunk_input: Optional[bool] = False,
sampling_rate: Optional[int] = None,
chunk_length: Optional[int] = 1,
**kwargs,
):
def get_audio_placeholder(audio_lens, chunk_input):
pool_step = 2 pool_step = 2
feature_lens = math.ceil(audio_lens / self.feature_extractor.hop_length) feature_lens = math.ceil(audio_lens / self.feature_extractor.hop_length)
@ -136,6 +127,15 @@ class MiniCPMOProcessor(ProcessorMixin):
return audio_placeholder return audio_placeholder
def audio_feature_extract(
self,
audios: Union[np.ndarray, List[np.ndarray], List[List[np.ndarray]]],
audio_parts: Optional[list] = None,
chunk_input: Optional[bool] = False,
sampling_rate: Optional[int] = None,
chunk_length: Optional[int] = 1,
**kwargs,
):
if isinstance(audios, np.ndarray): if isinstance(audios, np.ndarray):
audios_list = [[audios]] audios_list = [[audios]]
elif isinstance(audios[0], np.ndarray): elif isinstance(audios[0], np.ndarray):
@ -156,7 +156,7 @@ class MiniCPMOProcessor(ProcessorMixin):
# audio placeholder not dependent on audio_parts # audio placeholder not dependent on audio_parts
for audios in audios_list: for audios in audios_list:
if audios: if audios:
audio_ph_list.append([get_audio_placeholder(len(a), chunk_input) for a in audios]) audio_ph_list.append([self.get_audio_placeholder(len(a), chunk_input, chunk_length) for a in audios])
else: else:
audio_ph_list.append([]) audio_ph_list.append([])