Skip to content

Update device-agnostic code on tutorials #4044

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
wants to merge 9 commits into
base: main
Choose a base branch
from
Open

Update device-agnostic code on tutorials #4044

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
9 commits
Select commit Hold shift + click to select a range
677778c
Updated device-agnostic code on tutorials
Aug 13, 2025
1a0b0d6
Rolled back changes
Aug 13, 2025
ea34618
Updated device-agnostic code on tutorials
Aug 13, 2025
11ac548
Removing device: torch.device
Aug 14, 2025
440de59
Merge pull request #1 from gera-aldama/gda/w2v
gera-aldama Aug 14, 2025
1d23777
Refactored Audio Resampling Tutorial to work with accelerator API (#2)
gera-aldama Sep 1, 2025
0a86fa4
Rolled back changes
Sep 1, 2025
f935fe5
Rolled back changes
Sep 1, 2025
04eeaaa
Rolled back changes
Sep 1, 2025
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
32 changes: 20 additions & 12 deletions examples/tutorials/audio_resampling_tutorial.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -16,6 +16,13 @@
print(torch.__version__)
print(torchaudio.__version__)

if torch.accelerator.is_available():
acc = torch.accelerator.current_accelerator()
device = torch.device(acc)
else:
device = torch.device("cpu")
print(f"Using device: {device}")

######################################################################
# Preparation
# -----------
Expand All @@ -41,7 +48,7 @@ def _get_log_freq(sample_rate, max_sweep_rate, offset):

"""
start, stop = math.log(offset), math.log(offset + max_sweep_rate // 2)
return torch.exp(torch.linspace(start, stop, sample_rate, dtype=torch.double)) - offset
return torch.exp(torch.linspace(start, stop, sample_rate, dtype=torch.double, device=device)) - offset


def _get_inverse_log_freq(freq, sample_rate, offset):
Expand Down Expand Up @@ -91,6 +98,7 @@ def plot_sweep(
freq_y = [f for f in freq if f in y_ticks and 1000 <= f <= sample_rate // 2]

figure, axis = plt.subplots(1, 1)
waveform = waveform.cpu() if torch.is_tensor(waveform) else waveform
_, _, _, cax = axis.specgram(waveform[0].numpy(), Fs=sample_rate)
plt.xticks(time, freq_x)
plt.yticks(freq_y, freq_y)
Expand Down Expand Up @@ -144,7 +152,7 @@ def plot_sweep(
waveform = get_sine_sweep(sample_rate)

plot_sweep(waveform, sample_rate, title="Original Waveform")
Audio(waveform.numpy()[0], rate=sample_rate)
Audio(waveform.cpu().numpy()[0], rate=sample_rate)

######################################################################
#
Expand All @@ -157,11 +165,11 @@ def plot_sweep(
# an explanation of how it happens, and why it looks like a reflection.

resample_rate = 32000
resampler = T.Resample(sample_rate, resample_rate, dtype=waveform.dtype)
resampler = T.Resample(sample_rate, resample_rate, dtype=waveform.dtype).to(device)
resampled_waveform = resampler(waveform)

plot_sweep(resampled_waveform, resample_rate, title="Resampled Waveform")
Audio(resampled_waveform.numpy()[0], rate=resample_rate)
Audio(resampled_waveform.cpu().numpy()[0], rate=resample_rate)

######################################################################
# Controling resampling quality with parameters
Expand All @@ -182,13 +190,13 @@ def plot_sweep(
sample_rate = 48000
resample_rate = 32000

resampled_waveform = F.resample(waveform, sample_rate, resample_rate, lowpass_filter_width=6)
resampled_waveform = F.resample(waveform, sample_rate, resample_rate, lowpass_filter_width=6).to(device)
plot_sweep(resampled_waveform, resample_rate, title="lowpass_filter_width=6")

######################################################################
#

resampled_waveform = F.resample(waveform, sample_rate, resample_rate, lowpass_filter_width=128)
resampled_waveform = F.resample(waveform, sample_rate, resample_rate, lowpass_filter_width=128).to(device)
plot_sweep(resampled_waveform, resample_rate, title="lowpass_filter_width=128")

######################################################################
Expand All @@ -208,13 +216,13 @@ def plot_sweep(
sample_rate = 48000
resample_rate = 32000

resampled_waveform = F.resample(waveform, sample_rate, resample_rate, rolloff=0.99)
resampled_waveform = F.resample(waveform, sample_rate, resample_rate, rolloff=0.99).to(device)
plot_sweep(resampled_waveform, resample_rate, title="rolloff=0.99")

######################################################################
#

resampled_waveform = F.resample(waveform, sample_rate, resample_rate, rolloff=0.8)
resampled_waveform = F.resample(waveform, sample_rate, resample_rate, rolloff=0.8).to(device)
plot_sweep(resampled_waveform, resample_rate, title="rolloff=0.8")


Expand All @@ -234,13 +242,13 @@ def plot_sweep(
sample_rate = 48000
resample_rate = 32000

resampled_waveform = F.resample(waveform, sample_rate, resample_rate, resampling_method="sinc_interp_hann")
resampled_waveform = F.resample(waveform, sample_rate, resample_rate, resampling_method="sinc_interp_hann").to(device)
plot_sweep(resampled_waveform, resample_rate, title="Hann Window Default")

######################################################################
#

resampled_waveform = F.resample(waveform, sample_rate, resample_rate, resampling_method="sinc_interp_kaiser")
resampled_waveform = F.resample(waveform, sample_rate, resample_rate, resampling_method="sinc_interp_kaiser").to(device)
plot_sweep(resampled_waveform, resample_rate, title="Kaiser Window Default")


Expand All @@ -267,7 +275,7 @@ def plot_sweep(
rolloff=0.9475937167399596,
resampling_method="sinc_interp_kaiser",
beta=14.769656459379492,
)
).to(device)
plot_sweep(resampled_waveform, resample_rate, title="Kaiser Window Best (torchaudio)")

######################################################################
Expand All @@ -282,7 +290,7 @@ def plot_sweep(
rolloff=0.85,
resampling_method="sinc_interp_kaiser",
beta=8.555504641634386,
)
).to(device)
plot_sweep(resampled_waveform, resample_rate, title="Kaiser Window Fast (torchaudio)")

######################################################################
Expand Down
8 changes: 7 additions & 1 deletion examples/tutorials/forced_alignment_tutorial.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -47,7 +47,13 @@
print(torchaudio.__version__)


device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
if torch.accelerator.is_available():
acc = torch.accelerator.current_accelerator()
device = torch.device(acc)
backend = torch.distributed.get_default_backend_for_device(device)
else:
device = torch.device("cpu")
backend = "gloo"
print(device)


Expand Down
8 changes: 7 additions & 1 deletion examples/tutorials/speech_recognition_pipeline_tutorial.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -42,7 +42,13 @@
print(torchaudio.__version__)

torch.random.manual_seed(0)
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
if torch.accelerator.is_available():
acc = torch.accelerator.current_accelerator()
device = torch.device(acc)
backend = torch.distributed.get_default_backend_for_device(device)
else:
device = torch.device("cpu")
backend = "gloo"

print(device)

Expand Down