Skip to content

fix: use persistent=False for recomputable buffers in transforms #4188

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
prisha2410 wants to merge 1 commit into
base: main
Choose a base branch
from
+12 −12
Open

fix: use persistent=False for recomputable buffers in transforms #4188

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
24 changes: 12 additions & 12 deletions src/torchaudio/transforms/_transforms.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -84,7 +84,7 @@ def __init__(
self.win_length = win_length if win_length is not None else n_fft
self.hop_length = hop_length if hop_length is not None else self.win_length // 2
window = window_fn(self.win_length) if wkwargs is None else window_fn(self.win_length, **wkwargs)
self.register_buffer("window", window)
self.register_buffer("window", window, persistent=False)
self.pad = pad
self.power = power
self.normalized = normalized
Expand Down Expand Up @@ -178,7 +178,7 @@ def __init__(
self.win_length = win_length if win_length is not None else n_fft
self.hop_length = hop_length if hop_length is not None else self.win_length // 2
window = window_fn(self.win_length) if wkwargs is None else window_fn(self.win_length, **wkwargs)
self.register_buffer("window", window)
self.register_buffer("window", window, persistent=False)
self.pad = pad
self.normalized = normalized
self.center = center
Expand Down Expand Up @@ -267,7 +267,7 @@ def __init__(
self.win_length = win_length if win_length is not None else n_fft
self.hop_length = hop_length if hop_length is not None else self.win_length // 2
window = window_fn(self.win_length) if wkwargs is None else window_fn(self.win_length, **wkwargs)
self.register_buffer("window", window)
self.register_buffer("window", window, persistent=False)
self.length = length
self.power = power
self.momentum = momentum
Expand Down Expand Up @@ -402,7 +402,7 @@ def __init__(
fb = F.melscale_fbanks(
self.n_stft, self.f_min, self.f_max, self.n_mels, self.sample_rate, self.norm, self.mel_scale
)
self.register_buffer("fb", fb)
self.register_buffer("fb", fb, persistent=False)

def forward(self, specgram: Tensor) -> Tensor:
r"""
Expand Down Expand Up @@ -486,7 +486,7 @@ def __init__(
fb = F.melscale_fbanks(
self.n_stft, self.f_min, self.f_max, self.n_mels, self.sample_rate, self.norm, self.mel_scale
)
self.register_buffer("fb", fb)
self.register_buffer("fb", fb, persistent=False)

def forward(self, melspec: Tensor) -> Tensor:
r"""
Expand Down Expand Up @@ -695,7 +695,7 @@ def __init__(
if self.n_mfcc > self.MelSpectrogram.n_mels:
raise ValueError("Cannot select more MFCC coefficients than # mel bins")
dct_mat = F.create_dct(self.n_mfcc, self.MelSpectrogram.n_mels, self.norm)
self.register_buffer("dct_mat", dct_mat)
self.register_buffer("dct_mat", dct_mat, persistent=False)
self.log_mels = log_mels

def forward(self, waveform: Tensor) -> Tensor:
Expand Down Expand Up @@ -798,10 +798,10 @@ def __init__(
n_filter=self.n_filter,
sample_rate=self.sample_rate,
)
self.register_buffer("filter_mat", filter_mat)
self.register_buffer("filter_mat", filter_mat, persistent=False)

dct_mat = F.create_dct(self.n_lfcc, self.n_filter, self.norm)
self.register_buffer("dct_mat", dct_mat)
self.register_buffer("dct_mat", dct_mat, persistent=False)
self.log_lf = log_lf

def forward(self, waveform: Tensor) -> Tensor:
Expand Down Expand Up @@ -974,7 +974,7 @@ def __init__(
beta,
dtype=dtype,
)
self.register_buffer("kernel", kernel)
self.register_buffer("kernel", kernel, persistent=False)

def forward(self, waveform: Tensor) -> Tensor:
r"""
Expand Down Expand Up @@ -1061,7 +1061,7 @@ def __init__(self, hop_length: Optional[int] = None, n_freq: int = 201, fixed_ra

n_fft = (n_freq - 1) * 2
hop_length = hop_length if hop_length is not None else n_fft // 2
self.register_buffer("phase_advance", torch.linspace(0, math.pi * hop_length, n_freq)[..., None])
self.register_buffer("phase_advance", torch.linspace(0, math.pi * hop_length, n_freq)[..., None], persistent=False)

def forward(self, complex_specgrams: Tensor, overriding_rate: Optional[float] = None) -> Tensor:
r"""
Expand Down Expand Up @@ -1663,7 +1663,7 @@ def __init__(
self.win_length = win_length if win_length is not None else n_fft
self.hop_length = hop_length if hop_length is not None else self.win_length // 2
window = window_fn(self.win_length) if wkwargs is None else window_fn(self.win_length, **wkwargs)
self.register_buffer("window", window)
self.register_buffer("window", window, persistent=False)
self.pad = pad

def forward(self, waveform: Tensor) -> Tensor:
Expand Down Expand Up @@ -1728,7 +1728,7 @@ def __init__(
self.win_length = win_length if win_length is not None else n_fft
self.hop_length = hop_length if hop_length is not None else self.win_length // 4
window = window_fn(self.win_length) if wkwargs is None else window_fn(self.win_length, **wkwargs)
self.register_buffer("window", window)
self.register_buffer("window", window, persistent=False)
rate = 2.0 ** (-float(n_steps) / bins_per_octave)
self.orig_freq = int(sample_rate / rate)
self.gcd = math.gcd(int(self.orig_freq), int(sample_rate))
Expand Down