joel/freqsplit
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

change model

Browse Source
This commit is contained in:
Joel Mathew Thomas
2025-01-28 22:39:45 +05:30
parent 2982aaaafd
commit 88ebfb66f6
1 changed files with 2 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/separation/convtasnet_wrapper.py
+2 -2
View File
@@ -1,13 +1,13 @@
import torch import torch
from asteroid.models import ConvTasNet from asteroid.models import ConvTasNet
def separate(audio, model_name='mpariente/ConvTasNet_WHAMR_enhsingle'): def separate(audio, model_name='mpariente/ConvTasNet_WHAM!_sepclean'):
""" """
Separates audio into sources using a pretrained Asteroid model. Separates audio into sources using a pretrained Asteroid model.
Args: Args:
audio (numpy.ndarray): The audio time series (1D numpy array). audio (numpy.ndarray): The audio time series (1D numpy array).
model_name (str): Name of the pretrained model from Asteroid. Default is 'mpariente/ConvTasNet_WHAMR_enhsingle'. model_name (str): Name of the pretrained model from Asteroid. Default is 'mpariente/ConvTasNet_WHAM!_sepclean'.
Returns: Returns:
list: List of separated sources as numpy arrays. list: List of separated sources as numpy arrays.