P/j
d dl mZ d dlmZ d dlmZ d dlZd dlZd dlm Z m
Z
e G d d Z e d ee d
d Z d
e _
e d ee
g d d Zde_
e d ee
g d d Zde_
y) ) dataclass)partial)CallableN)conv_tasnet_base hdemucs_highc e Zd ZU dZeed< eg ej j f ed< e
ed< e de
fd Z dej j fdZ
y) SourceSeparationBundleu Dataclass that bundles components for performing source separation.
Example
>>> import torchaudio
>>> from torchaudio.pipelines import CONVTASNET_BASE_LIBRI2MIX
>>> import torch
>>>
>>> # Build the separation model.
>>> model = CONVTASNET_BASE_LIBRI2MIX.get_model()
>>> 100%|███████████████████████████████|19.1M/19.1M [00:04<00:00, 4.93MB/s]
>>>
>>> # Instantiate the test set of Libri2Mix dataset.
>>> dataset = torchaudio.datasets.LibriMix("/home/datasets/", subset="test")
>>>
>>> # Apply source separation on mixture audio.
>>> for i, data in enumerate(dataset):
>>> sample_rate, mixture, clean_sources = data
>>> # Make sure the shape of input suits the model requirement.
>>> mixture = mixture.reshape(1, 1, -1)
>>> estimated_sources = model(mixture)
>>> score = si_snr_pit(estimated_sources, clean_sources) # for demonstration
>>> print(f"Si-SNR score is : {score}.)
>>> break
>>> Si-SNR score is : 16.24.
>>>
_model_path_model_factory_func _sample_ratereturnc | j S )zSSample rate of the audio that the model is trained on.
:type: int
)r )selfs s/root/repos/pocs/video-evals/.venv/lib/python3.12/site-packages/torchaudio/pipelines/_source_separation_pipeline.py sample_ratez"SourceSeparationBundle.sample_rate, s c | j }t j j | j }t j | }|j | |j |S )z3Construct the model and load the pretrained weight.) r
torchaudioutilsdownload_assetr
torchloadload_state_dicteval)r modelpath
state_dicts r get_modelz SourceSeparationBundle.get_model4 sV ((* ..t/?/?@ZZ%
j)
r N)__name__
__module__ __qualname____doc__str__annotations__r r nnModuleintpropertyr r r r r r sZ 6 !"ehhoo"566
!S ! !588?? r r z$models/conv_tasnet_base_libri2mix.pt ) num_sourcesi@ )r
r r a Pre-trained Source Separation pipeline with *ConvTasNet*
:cite:`Luo_2019` trained on *Libri2Mix dataset* :cite:`cosentino2020librimix`.
The source separation model is constructed by :func:`~torchaudio.models.conv_tasnet_base`
and is trained using the training script ``lightning_train.py``
`here `__
with default arguments.
Please refer to :class:`SourceSeparationBundle` for usage instructions.
z models/hdemucs_high_trained.pt)drumsbassothervocals)sourcesiD a Pre-trained music source separation pipeline with
*Hybrid Demucs* :cite:`defossez2021hybrid` trained on both training and test sets of
MUSDB-HQ :cite:`MUSDB18HQ` and an additional 150 extra songs from an internal database
that was specifically produced for Meta.
The model is constructed by :func:`~torchaudio.models.hdemucs_high`.
Training was performed in the original HDemucs repository `here `__.
Please refer to :class:`SourceSeparationBundle` for usage instructions.
z#models/hdemucs_high_musdbhq_only.pta Pre-trained music source separation pipeline with
*Hybrid Demucs* :cite:`defossez2021hybrid` trained on the training set of MUSDB-HQ :cite:`MUSDB18HQ`.
The model is constructed by :func:`~torchaudio.models.hdemucs_high`.
Training was performed in the original HDemucs repository `here `__.
Please refer to :class:`SourceSeparationBundle` for usage instructions.
) dataclassesr functoolsr typingr r r torchaudio.modelsr r r CONVTASNET_BASE_LIBRI2MIXr" HDEMUCS_HIGH_MUSDB_PLUSHDEMUCS_HIGH_MUSDBr) r r r8 s ! < / / /d 36 0a@
% ! 10 6Z[
# ,5 6Z[
r