deepgeek/MuseTalk
2
0
Fork 0
mirror of https://github.com/TMElyralab/MuseTalk.git synced 2026-02-04 17:39:20 +08:00
Code Issues Packages Projects Releases Wiki Activity

feat: real-time infer (#286)

Browse Source 
* feat: realtime infer

* cchore: infer script
This commit is contained in:
Zhizhou Zhong
2025-04-02 19:13:18 +08:00
committed by GitHub
parent fbe6a97dff
commit 39ccf69f36
11 changed files with 490 additions and 592 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
musetalk/utils/audio_processor.py
View File

@@ -11,7 +11,7 @@ class AudioProcessor:
def __init__(self, feature_extractor_path="openai/whisper-tiny/"):
self.feature_extractor = AutoFeatureExtractor.from_pretrained(feature_extractor_path)
def get_audio_feature(self, wav_path, start_index=0):
def get_audio_feature(self, wav_path, start_index=0, weight_dtype=None):
if not os.path.exists(wav_path):
return None
librosa_output, sampling_rate = librosa.load(wav_path, sr=16000)
@@ -27,6 +27,8 @@ class AudioProcessor:
return_tensors="pt",
sampling_rate=sampling_rate
).input_features
if weight_dtype is not None:
audio_feature = audio_feature.to(dtype=weight_dtype)
features.append(audio_feature)
return features, len(librosa_output)