Commit 95b8cd23 authored by Phil Wang's avatar Phil Wang
Browse files

oops

parent b6e5af78

audiolm_pytorch/audiolm_pytorch.py

+5 −5
Original line number Diff line number Diff line
@@ -304,10 +304,10 @@ class SoundStream(nn.Module): 
        feature_loss_weight = 100,

        quantize_dropout = True,

        quantize_dropout_cutoff_index = 0,

        target_sample_khz = 24000

        target_sample_hz = 24000

    ):

        super().__init__()

        self.target_sample_khz = target_sample_khz # for resampling on the fly

        self.target_sample_hz = target_sample_hz # for resampling on the fly



        self.single_channel = input_channels == 1

        self.strides = strides
@@ -375,10 +375,10 @@ class SoundStream(nn.Module): 
        return_discr_loss = False,

        return_discr_losses_separately = False,

        return_recons_only = False,

        input_sample_khz = None

        input_sample_hz = None

    ):

        if exists(input_sample_khz):

            x = resample(x, input_sample_khz, self.target_sample_khz)

        if exists(input_sample_hz):

            x = resample(x, input_sample_hz, self.target_sample_hz)



        x = curtail_to_multiple(x, self.seq_len_multiple_of)

audiolm_pytorch/data.py

+5 −5
Original line number Diff line number Diff line
@@ -21,7 +21,7 @@ class SoundDataset(Dataset): 
        folder,

        exts = ['flac', 'wav'],

        max_length = None,

        target_sample_khz = None,

        target_sample_hz = None,

        seq_len_multiple_of = None

    ):

        super().__init__()
@@ -34,7 +34,7 @@ class SoundDataset(Dataset): 
        self.files = files

        self.max_length = max_length



        self.target_sample_khz = target_sample_khz

        self.target_sample_hz = target_sample_hz

        self.seq_len_multiple_of = seq_len_multiple_of



    def __len__(self):
@@ -42,12 +42,12 @@ class SoundDataset(Dataset): 


    def __getitem__(self, idx):

        file = self.files[idx]

        data, sample_khz = torchaudio.load(file)

        data, sample_hz = torchaudio.load(file)



        data = rearrange(data, '1 ... -> ...')



        if exists(self.target_sample_khz):

            data = torchaudio.functional.resample(data, sample_khz, self.target_sample_khz)

        if exists(self.target_sample_hz):

            data = torchaudio.functional.resample(data, sample_hz, self.target_sample_hz)



        if exists(self.max_length):

            data = data[:self.max_length]

audiolm_pytorch/hubert_kmeans.py

+5 −5
Original line number Diff line number Diff line
@@ -19,11 +19,11 @@ class HubertWithKmeans(nn.Module): 
        self,

        checkpoint_path,

        kmeans_path,

        target_sample_khz = 50000,

        target_sample_hz = 50000,

        seq_len_multiple_of = None

    ):

        super().__init__()

        self.target_sample_khz = target_sample_khz

        self.target_sample_hz = target_sample_hz

        self.seq_len_multiple_of = seq_len_multiple_of



        model_path = Path(checkpoint_path)
@@ -55,12 +55,12 @@ class HubertWithKmeans(nn.Module): 
        self,

        wav_input,

        flatten = True,

        input_sample_khz = None

        input_sample_hz = None

    ):

        device = wav_input.device



        if exists(input_sample_khz):

            wav_input = resample(wav_input, input_sample_khz, self.target_sample_khz)

        if exists(input_sample_hz):

            wav_input = resample(wav_input, input_sample_hz, self.target_sample_hz)



        if exists(self.seq_len_multiple_of):

            wav_input = curtail_to_multiple(wav_input, self.seq_len_multiple_of)

audiolm_pytorch/vq_wav2vec.py

+5 −5
Original line number Diff line number Diff line
@@ -17,11 +17,11 @@ class FairseqVQWav2Vec(nn.Module): 
    def __init__(

        self,

        checkpoint_path,

        target_sample_khz = 24000,

        target_sample_hz = 24000,

        seq_len_multiple_of = None

    ):

        super().__init__()

        self.target_sample_khz = target_sample_khz

        self.target_sample_hz = target_sample_hz

        self.seq_len_multiple_of = seq_len_multiple_of



        path = Path(checkpoint_path)
@@ -47,10 +47,10 @@ class FairseqVQWav2Vec(nn.Module): 
        self,

        wav_input,

        flatten = True,

        input_sample_khz = None

        input_sample_hz = None

    ):

        if exists(input_sample_khz):

            wav_input = resample(wav_input, input_sample_khz, self.target_sample_khz)

        if exists(input_sample_hz):

            wav_input = resample(wav_input, input_sample_hz, self.target_sample_hz)



        if exists(self.seq_len_multiple_of):

            wav_input = curtail_to_multiple(wav_input, self.seq_len_multiple_of)

setup.py

+1 −1
Original line number Diff line number Diff line
@@ -3,7 +3,7 @@ from setuptools import setup, find_packages 
setup(

  name = 'audiolm-pytorch',

  packages = find_packages(exclude=[]),

  version = '0.0.24',

  version = '0.0.25',

  license='MIT',

  description = 'AudioLM - Language Modeling Approach to Audio Generation from Google Research - Pytorch',

  author = 'Phil Wang',