Spaces:
Runtime error
Runtime error
different seed for latent and denoisng
Browse files- audiodiffusion/__init__.py +50 -24
audiodiffusion/__init__.py
CHANGED
@@ -59,37 +59,44 @@ class AudioDiffusion:
|
|
59 |
top_db=top_db)
|
60 |
|
61 |
def generate_spectrogram_and_audio(
|
62 |
-
|
63 |
-
|
64 |
-
|
65 |
-
|
|
|
66 |
"""Generate random mel spectrogram and convert to audio.
|
67 |
|
68 |
Args:
|
69 |
steps (int): number of de-noising steps to perform (defaults to num_train_timesteps)
|
70 |
generator (torch.Generator): random number generator or None
|
|
|
|
|
71 |
|
72 |
Returns:
|
73 |
PIL Image: mel spectrogram
|
74 |
(float, np.ndarray): sample rate and raw audio
|
75 |
"""
|
76 |
-
images, (sample_rate,
|
77 |
-
|
78 |
-
|
79 |
-
|
|
|
|
|
|
|
80 |
return images[0], (sample_rate, audios[0])
|
81 |
|
82 |
def generate_spectrogram_and_audio_from_audio(
|
83 |
-
|
84 |
-
|
85 |
-
|
86 |
-
|
87 |
-
|
88 |
-
|
89 |
-
|
90 |
-
|
91 |
-
|
92 |
-
|
|
|
93 |
"""Generate random mel spectrogram from audio input and convert to audio.
|
94 |
|
95 |
Args:
|
@@ -101,6 +108,8 @@ class AudioDiffusion:
|
|
101 |
generator (torch.Generator): random number generator or None
|
102 |
mask_start_secs (float): number of seconds of audio to mask (not generate) at start
|
103 |
mask_end_secs (float): number of seconds of audio to mask (not generate) at end
|
|
|
|
|
104 |
|
105 |
Returns:
|
106 |
PIL Image: mel spectrogram
|
@@ -117,7 +126,9 @@ class AudioDiffusion:
|
|
117 |
steps=steps,
|
118 |
generator=generator,
|
119 |
mask_start_secs=mask_start_secs,
|
120 |
-
mask_end_secs=mask_end_secs
|
|
|
|
|
121 |
return images[0], (sample_rate, audios[0])
|
122 |
|
123 |
@staticmethod
|
@@ -160,7 +171,9 @@ class AudioDiffusionPipeline(DiffusionPipeline):
|
|
160 |
steps: int = 1000,
|
161 |
generator: torch.Generator = None,
|
162 |
mask_start_secs: float = 0,
|
163 |
-
mask_end_secs: float = 0
|
|
|
|
|
164 |
) -> Tuple[List[Image.Image], Tuple[int, List[np.ndarray]]]:
|
165 |
"""Generate random mel spectrogram from audio input and convert to audio.
|
166 |
|
@@ -175,6 +188,8 @@ class AudioDiffusionPipeline(DiffusionPipeline):
|
|
175 |
generator (torch.Generator): random number generator or None
|
176 |
mask_start_secs (float): number of seconds of audio to mask (not generate) at start
|
177 |
mask_end_secs (float): number of seconds of audio to mask (not generate) at end
|
|
|
|
|
178 |
|
179 |
Returns:
|
180 |
List[PIL Image]: mel spectrograms
|
@@ -182,6 +197,7 @@ class AudioDiffusionPipeline(DiffusionPipeline):
|
|
182 |
"""
|
183 |
|
184 |
self.scheduler.set_timesteps(steps)
|
|
|
185 |
mask = None
|
186 |
images = noise = torch.randn(
|
187 |
(batch_size, self.unet.in_channels, mel.y_res, mel.x_res),
|
@@ -218,10 +234,20 @@ class AudioDiffusionPipeline(DiffusionPipeline):
|
|
218 |
for step, t in enumerate(
|
219 |
self.progress_bar(self.scheduler.timesteps[start_step:])):
|
220 |
model_output = self.unet(images, t)['sample']
|
221 |
-
|
222 |
-
|
223 |
-
|
224 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
225 |
|
226 |
if mask is not None:
|
227 |
if mask_start > 0:
|
|
|
59 |
top_db=top_db)
|
60 |
|
61 |
def generate_spectrogram_and_audio(
|
62 |
+
self,
|
63 |
+
steps: int = 1000,
|
64 |
+
generator: torch.Generator = None,
|
65 |
+
step_generator: torch.Generator = None,
|
66 |
+
eta: float = 0) -> Tuple[Image.Image, Tuple[int, np.ndarray]]:
|
67 |
"""Generate random mel spectrogram and convert to audio.
|
68 |
|
69 |
Args:
|
70 |
steps (int): number of de-noising steps to perform (defaults to num_train_timesteps)
|
71 |
generator (torch.Generator): random number generator or None
|
72 |
+
step_generator (torch.Generator): random number generator used to denoise or None
|
73 |
+
eta (float): parameter between 0 and 1 used with DDIM scheduler
|
74 |
|
75 |
Returns:
|
76 |
PIL Image: mel spectrogram
|
77 |
(float, np.ndarray): sample rate and raw audio
|
78 |
"""
|
79 |
+
images, (sample_rate,
|
80 |
+
audios) = self.pipe(mel=self.mel,
|
81 |
+
batch_size=1,
|
82 |
+
steps=steps,
|
83 |
+
generator=generator,
|
84 |
+
step_generator=step_generator,
|
85 |
+
eta=eta)
|
86 |
return images[0], (sample_rate, audios[0])
|
87 |
|
88 |
def generate_spectrogram_and_audio_from_audio(
|
89 |
+
self,
|
90 |
+
audio_file: str = None,
|
91 |
+
raw_audio: np.ndarray = None,
|
92 |
+
slice: int = 0,
|
93 |
+
start_step: int = 0,
|
94 |
+
steps: int = 1000,
|
95 |
+
generator: torch.Generator = None,
|
96 |
+
mask_start_secs: float = 0,
|
97 |
+
mask_end_secs: float = 0,
|
98 |
+
step_generator: torch.Generator = None,
|
99 |
+
eta: float = 0) -> Tuple[Image.Image, Tuple[int, np.ndarray]]:
|
100 |
"""Generate random mel spectrogram from audio input and convert to audio.
|
101 |
|
102 |
Args:
|
|
|
108 |
generator (torch.Generator): random number generator or None
|
109 |
mask_start_secs (float): number of seconds of audio to mask (not generate) at start
|
110 |
mask_end_secs (float): number of seconds of audio to mask (not generate) at end
|
111 |
+
step_generator (torch.Generator): random number generator used to denoise or None
|
112 |
+
eta (float): parameter between 0 and 1 used with DDIM scheduler
|
113 |
|
114 |
Returns:
|
115 |
PIL Image: mel spectrogram
|
|
|
126 |
steps=steps,
|
127 |
generator=generator,
|
128 |
mask_start_secs=mask_start_secs,
|
129 |
+
mask_end_secs=mask_end_secs,
|
130 |
+
step_generator=step_generator,
|
131 |
+
eta=eta)
|
132 |
return images[0], (sample_rate, audios[0])
|
133 |
|
134 |
@staticmethod
|
|
|
171 |
steps: int = 1000,
|
172 |
generator: torch.Generator = None,
|
173 |
mask_start_secs: float = 0,
|
174 |
+
mask_end_secs: float = 0,
|
175 |
+
step_generator: torch.Generator = None,
|
176 |
+
eta: float = 0
|
177 |
) -> Tuple[List[Image.Image], Tuple[int, List[np.ndarray]]]:
|
178 |
"""Generate random mel spectrogram from audio input and convert to audio.
|
179 |
|
|
|
188 |
generator (torch.Generator): random number generator or None
|
189 |
mask_start_secs (float): number of seconds of audio to mask (not generate) at start
|
190 |
mask_end_secs (float): number of seconds of audio to mask (not generate) at end
|
191 |
+
step_generator (torch.Generator): random number generator used to denoise or None
|
192 |
+
eta (float): parameter between 0 and 1 used with DDIM scheduler
|
193 |
|
194 |
Returns:
|
195 |
List[PIL Image]: mel spectrograms
|
|
|
197 |
"""
|
198 |
|
199 |
self.scheduler.set_timesteps(steps)
|
200 |
+
step_generator = step_generator or generator
|
201 |
mask = None
|
202 |
images = noise = torch.randn(
|
203 |
(batch_size, self.unet.in_channels, mel.y_res, mel.x_res),
|
|
|
234 |
for step, t in enumerate(
|
235 |
self.progress_bar(self.scheduler.timesteps[start_step:])):
|
236 |
model_output = self.unet(images, t)['sample']
|
237 |
+
|
238 |
+
if isinstance(self.scheduler, DDIMScheduler):
|
239 |
+
images = self.scheduler.step(
|
240 |
+
model_output=model_output,
|
241 |
+
timestep=t,
|
242 |
+
sample=images,
|
243 |
+
eta=eta,
|
244 |
+
generator=step_generator)['prev_sample']
|
245 |
+
else:
|
246 |
+
images = self.scheduler.step(
|
247 |
+
model_output=model_output,
|
248 |
+
timestep=t,
|
249 |
+
sample=images,
|
250 |
+
generator=step_generator)['prev_sample']
|
251 |
|
252 |
if mask is not None:
|
253 |
if mask_start > 0:
|