Spaces:
Running
on
Zero
Running
on
Zero
update layout
Browse files
app.py
CHANGED
@@ -50,11 +50,43 @@ def infer(
|
|
50 |
):
|
51 |
device = "cuda" if torch.cuda.is_available() else "cpu"
|
52 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
53 |
if randomize_seed:
|
54 |
seed = random.randint(0, MAX_SEED)
|
55 |
prompts = [prompt1, prompt2]
|
56 |
generator = torch.Generator().manual_seed(seed)
|
57 |
-
|
58 |
interpolated_prompt_embeds, prompt_metadata = synth.interpolatePrompts(
|
59 |
prompts,
|
60 |
pipe,
|
@@ -65,18 +97,16 @@ def infer(
|
|
65 |
)
|
66 |
negative_prompts = [negative_prompt, negative_prompt]
|
67 |
if negative_prompts != ["", ""]:
|
68 |
-
interpolated_negative_prompts_embeds,
|
69 |
-
|
70 |
-
|
71 |
-
|
72 |
-
|
73 |
-
|
74 |
-
|
75 |
-
device=device,
|
76 |
-
)
|
77 |
)
|
78 |
else:
|
79 |
-
interpolated_negative_prompts_embeds,
|
80 |
interpolated_prompt_embeds
|
81 |
), None
|
82 |
|
@@ -129,7 +159,7 @@ def infer(
|
|
129 |
* 100
|
130 |
)
|
131 |
|
132 |
-
return image, seed, ssim_score, cosine_sim
|
133 |
|
134 |
|
135 |
examples1 = [
|
@@ -141,12 +171,6 @@ examples2 = [
|
|
141 |
"A photo of a beagle, a type of dog",
|
142 |
]
|
143 |
|
144 |
-
css = """
|
145 |
-
#col-container {
|
146 |
-
margin: 0 auto;
|
147 |
-
}
|
148 |
-
"""
|
149 |
-
|
150 |
|
151 |
def update_steps(total_steps, interpolation_step):
|
152 |
if interpolation_step > total_steps:
|
@@ -154,181 +178,191 @@ def update_steps(total_steps, interpolation_step):
|
|
154 |
return gr.update(maximum=total_steps // 2)
|
155 |
|
156 |
|
|
|
|
|
|
|
|
|
|
|
|
|
157 |
if torch.cuda.is_available():
|
158 |
power_device = "GPU"
|
159 |
else:
|
160 |
power_device = "CPU"
|
161 |
|
162 |
-
with gr.Blocks(
|
163 |
|
164 |
-
|
165 |
-
gr.Markdown(
|
166 |
-
f"""
|
167 |
-
# Data Augmentation with Image-to-Image Diffusion Models via Prompt Interpolation
|
168 |
-
Currently running on {power_device}.
|
169 |
"""
|
170 |
-
|
171 |
-
|
172 |
-
|
173 |
-
|
174 |
-
with gr.
|
175 |
-
prompt1 = gr.Text(
|
176 |
-
label="Prompt 1",
|
177 |
-
show_label=True,
|
178 |
-
max_lines=1,
|
179 |
-
placeholder="Enter your first prompt",
|
180 |
-
container=False,
|
181 |
-
)
|
182 |
-
with gr.Row():
|
183 |
-
prompt2 = gr.Text(
|
184 |
-
label="Prompt 2",
|
185 |
-
show_label=True,
|
186 |
-
max_lines=1,
|
187 |
-
placeholder="Enter your second prompt",
|
188 |
-
container=False,
|
189 |
-
)
|
190 |
-
with gr.Row():
|
191 |
-
gr.Examples(
|
192 |
-
examples=examples1, inputs=[prompt1], label="Example for Prompt 1"
|
193 |
-
)
|
194 |
-
gr.Examples(
|
195 |
-
examples=examples2, inputs=[prompt2], label="Example for Prompt 2"
|
196 |
-
)
|
197 |
-
|
198 |
-
with gr.Row():
|
199 |
-
num_interpolation_steps = gr.Slider(
|
200 |
-
label="Total interpolation steps",
|
201 |
-
minimum=2,
|
202 |
-
maximum=32,
|
203 |
-
step=2,
|
204 |
-
value=16,
|
205 |
-
)
|
206 |
-
interpolation_step = gr.Slider(
|
207 |
-
label="Specific Interpolation Step",
|
208 |
-
minimum=1,
|
209 |
-
maximum=8,
|
210 |
-
step=1,
|
211 |
-
value=8,
|
212 |
-
)
|
213 |
-
num_interpolation_steps.change(
|
214 |
-
fn=update_steps,
|
215 |
-
inputs=[num_interpolation_steps, interpolation_step],
|
216 |
-
outputs=[interpolation_step],
|
217 |
-
)
|
218 |
-
run_button = gr.Button("Run", scale=0)
|
219 |
-
|
220 |
-
result = gr.Image(label="Result", show_label=False)
|
221 |
-
|
222 |
-
with gr.Accordion("Advanced Settings", open=True):
|
223 |
-
|
224 |
-
negative_prompt = gr.Text(
|
225 |
-
label="Negative prompt",
|
226 |
-
max_lines=1,
|
227 |
-
placeholder="Enter a negative prompt",
|
228 |
-
visible=False,
|
229 |
-
)
|
230 |
-
|
231 |
-
seed = gr.Slider(
|
232 |
-
label="Seed",
|
233 |
-
minimum=0,
|
234 |
-
maximum=MAX_SEED,
|
235 |
-
step=1,
|
236 |
-
value=0,
|
237 |
-
)
|
238 |
|
239 |
-
|
240 |
|
241 |
-
gr.Markdown("Negative Prompt: ")
|
242 |
with gr.Row():
|
243 |
-
|
244 |
-
label="
|
245 |
show_label=True,
|
246 |
max_lines=1,
|
247 |
-
|
248 |
container=False,
|
249 |
)
|
250 |
with gr.Row():
|
251 |
-
|
252 |
-
|
253 |
-
|
254 |
-
|
255 |
-
|
256 |
-
|
257 |
-
value=512,
|
258 |
-
)
|
259 |
-
|
260 |
-
height = gr.Slider(
|
261 |
-
label="Height",
|
262 |
-
minimum=256,
|
263 |
-
maximum=MAX_IMAGE_SIZE,
|
264 |
-
step=32,
|
265 |
-
value=512,
|
266 |
)
|
267 |
-
|
268 |
with gr.Row():
|
269 |
-
|
270 |
-
|
271 |
-
|
272 |
-
|
273 |
-
|
274 |
-
step=0.1,
|
275 |
-
value=8.0,
|
276 |
)
|
277 |
|
278 |
-
|
279 |
-
|
|
|
280 |
minimum=1,
|
281 |
-
maximum=
|
282 |
step=1,
|
283 |
-
value=
|
284 |
)
|
285 |
-
|
286 |
-
|
287 |
-
label="Number of sampling steps in the middle of interpolation",
|
288 |
minimum=2,
|
289 |
-
maximum=
|
290 |
step=2,
|
291 |
value=16,
|
292 |
)
|
293 |
-
|
294 |
-
|
295 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
296 |
minimum=0,
|
297 |
-
maximum=
|
298 |
-
step=
|
299 |
value=0,
|
300 |
)
|
301 |
-
|
302 |
-
|
303 |
-
|
304 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
305 |
)
|
306 |
-
with gr.Row():
|
307 |
-
show_seed = gr.Label(label="Seed:", value="Randomized seed")
|
308 |
-
ssim_score = gr.Label(label="SSIM Score:", value="Generate to see score")
|
309 |
-
cos_sim = gr.Label(label="CLIP Score:", value="Generate to see score")
|
310 |
-
run_button.click(
|
311 |
-
fn=infer,
|
312 |
-
inputs=[
|
313 |
-
input_image,
|
314 |
-
prompt1,
|
315 |
-
prompt2,
|
316 |
-
negative_prompt,
|
317 |
-
seed,
|
318 |
-
randomize_seed,
|
319 |
-
width,
|
320 |
-
height,
|
321 |
-
guidance_scale,
|
322 |
-
interpolation_step,
|
323 |
-
num_inference_steps,
|
324 |
-
num_interpolation_steps,
|
325 |
-
sample_mid_interpolation,
|
326 |
-
remove_n_middle,
|
327 |
-
],
|
328 |
-
outputs=[result, show_seed, ssim_score, cos_sim],
|
329 |
-
)
|
330 |
|
331 |
-
demo.queue().launch()
|
332 |
|
333 |
"""
|
334 |
input_image,
|
|
|
50 |
):
|
51 |
device = "cuda" if torch.cuda.is_available() else "cpu"
|
52 |
|
53 |
+
# Input Validation
|
54 |
+
try:
|
55 |
+
assert num_interpolation_steps % 2 == 0
|
56 |
+
except AssertionError:
|
57 |
+
raise ValueError("num_interpolation_steps must be an even number")
|
58 |
+
try:
|
59 |
+
assert sample_mid_interpolation % 2 == 0
|
60 |
+
except AssertionError:
|
61 |
+
raise ValueError("sample_mid_interpolation must be an even number")
|
62 |
+
try:
|
63 |
+
assert remove_n_middle % 2 == 0
|
64 |
+
except AssertionError:
|
65 |
+
raise ValueError("remove_n_middle must be an even number")
|
66 |
+
try:
|
67 |
+
assert num_interpolation_steps >= sample_mid_interpolation
|
68 |
+
except AssertionError:
|
69 |
+
raise ValueError(
|
70 |
+
"num_interpolation_steps must be greater than or equal to sample_mid_interpolation"
|
71 |
+
)
|
72 |
+
try:
|
73 |
+
assert num_interpolation_steps >= 2 and sample_mid_interpolation >= 2
|
74 |
+
except AssertionError:
|
75 |
+
raise ValueError(
|
76 |
+
"num_interpolation_steps and sample_mid_interpolation must be greater than or equal to 2"
|
77 |
+
)
|
78 |
+
try:
|
79 |
+
assert sample_mid_interpolation - remove_n_middle >= 2
|
80 |
+
except AssertionError:
|
81 |
+
raise ValueError(
|
82 |
+
"sample_mid_interpolation must be greater than or equal to remove_n_middle + 2"
|
83 |
+
)
|
84 |
+
|
85 |
if randomize_seed:
|
86 |
seed = random.randint(0, MAX_SEED)
|
87 |
prompts = [prompt1, prompt2]
|
88 |
generator = torch.Generator().manual_seed(seed)
|
89 |
+
|
90 |
interpolated_prompt_embeds, prompt_metadata = synth.interpolatePrompts(
|
91 |
prompts,
|
92 |
pipe,
|
|
|
97 |
)
|
98 |
negative_prompts = [negative_prompt, negative_prompt]
|
99 |
if negative_prompts != ["", ""]:
|
100 |
+
interpolated_negative_prompts_embeds, _ = synth.interpolatePrompts(
|
101 |
+
negative_prompts,
|
102 |
+
pipe,
|
103 |
+
num_interpolation_steps,
|
104 |
+
sample_mid_interpolation,
|
105 |
+
remove_n_middle=remove_n_middle,
|
106 |
+
device=device,
|
|
|
|
|
107 |
)
|
108 |
else:
|
109 |
+
interpolated_negative_prompts_embeds, _ = [None] * len(
|
110 |
interpolated_prompt_embeds
|
111 |
), None
|
112 |
|
|
|
159 |
* 100
|
160 |
)
|
161 |
|
162 |
+
return image, seed, round(ssim_score, 4), round(cosine_sim, 2)
|
163 |
|
164 |
|
165 |
examples1 = [
|
|
|
171 |
"A photo of a beagle, a type of dog",
|
172 |
]
|
173 |
|
|
|
|
|
|
|
|
|
|
|
|
|
174 |
|
175 |
def update_steps(total_steps, interpolation_step):
|
176 |
if interpolation_step > total_steps:
|
|
|
178 |
return gr.update(maximum=total_steps // 2)
|
179 |
|
180 |
|
181 |
+
def update_sampling_steps(total_steps, sample_steps):
|
182 |
+
# if sample_steps > total_steps:
|
183 |
+
# return gr.update(value=total_steps)
|
184 |
+
return gr.update(value=total_steps)
|
185 |
+
|
186 |
+
|
187 |
if torch.cuda.is_available():
|
188 |
power_device = "GPU"
|
189 |
else:
|
190 |
power_device = "CPU"
|
191 |
|
192 |
+
with gr.Blocks(title="Generative Date Augmentation") as demo:
|
193 |
|
194 |
+
gr.Markdown(
|
|
|
|
|
|
|
|
|
195 |
"""
|
196 |
+
# Data Augmentation with Image-to-Image Diffusion Models via Prompt Interpolation.
|
197 |
+
"""
|
198 |
+
)
|
199 |
+
with gr.Row():
|
200 |
+
with gr.Column():
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
201 |
|
202 |
+
input_image = gr.Image(type="pil", label="Image to Augment")
|
203 |
|
|
|
204 |
with gr.Row():
|
205 |
+
prompt1 = gr.Text(
|
206 |
+
label="Prompt 1",
|
207 |
show_label=True,
|
208 |
max_lines=1,
|
209 |
+
placeholder="Enter your first prompt",
|
210 |
container=False,
|
211 |
)
|
212 |
with gr.Row():
|
213 |
+
prompt2 = gr.Text(
|
214 |
+
label="Prompt 2",
|
215 |
+
show_label=True,
|
216 |
+
max_lines=1,
|
217 |
+
placeholder="Enter your second prompt",
|
218 |
+
container=False,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
219 |
)
|
|
|
220 |
with gr.Row():
|
221 |
+
gr.Examples(
|
222 |
+
examples=examples1, inputs=[prompt1], label="Example for Prompt 1"
|
223 |
+
)
|
224 |
+
gr.Examples(
|
225 |
+
examples=examples2, inputs=[prompt2], label="Example for Prompt 2"
|
|
|
|
|
226 |
)
|
227 |
|
228 |
+
with gr.Row():
|
229 |
+
interpolation_step = gr.Slider(
|
230 |
+
label="Specific Interpolation Step",
|
231 |
minimum=1,
|
232 |
+
maximum=8,
|
233 |
step=1,
|
234 |
+
value=8,
|
235 |
)
|
236 |
+
num_interpolation_steps = gr.Slider(
|
237 |
+
label="Total interpolation steps",
|
|
|
238 |
minimum=2,
|
239 |
+
maximum=32,
|
240 |
step=2,
|
241 |
value=16,
|
242 |
)
|
243 |
+
num_interpolation_steps.change(
|
244 |
+
fn=update_steps,
|
245 |
+
inputs=[num_interpolation_steps, interpolation_step],
|
246 |
+
outputs=[interpolation_step],
|
247 |
+
)
|
248 |
+
run_button = gr.Button("Run", scale=0)
|
249 |
+
with gr.Accordion("Advanced Settings", open=True):
|
250 |
+
negative_prompt = gr.Text(
|
251 |
+
label="Negative prompt",
|
252 |
+
max_lines=1,
|
253 |
+
placeholder="Enter a negative prompt",
|
254 |
+
visible=False,
|
255 |
+
)
|
256 |
+
seed = gr.Slider(
|
257 |
+
label="Seed",
|
258 |
minimum=0,
|
259 |
+
maximum=MAX_SEED,
|
260 |
+
step=1,
|
261 |
value=0,
|
262 |
)
|
263 |
+
randomize_seed = gr.Checkbox(label="Randomize seed", value=True)
|
264 |
+
gr.Markdown("Negative Prompt: ")
|
265 |
+
with gr.Row():
|
266 |
+
negative_prompt = gr.Text(
|
267 |
+
label="Negative Prompt",
|
268 |
+
show_label=True,
|
269 |
+
max_lines=1,
|
270 |
+
value="blurry image, disfigured, deformed, distorted, cartoon, drawings",
|
271 |
+
container=False,
|
272 |
+
)
|
273 |
+
with gr.Row():
|
274 |
+
width = gr.Slider(
|
275 |
+
label="Width",
|
276 |
+
minimum=256,
|
277 |
+
maximum=MAX_IMAGE_SIZE,
|
278 |
+
step=32,
|
279 |
+
value=512,
|
280 |
+
)
|
281 |
+
height = gr.Slider(
|
282 |
+
label="Height",
|
283 |
+
minimum=256,
|
284 |
+
maximum=MAX_IMAGE_SIZE,
|
285 |
+
step=32,
|
286 |
+
value=512,
|
287 |
+
)
|
288 |
+
with gr.Row():
|
289 |
+
guidance_scale = gr.Slider(
|
290 |
+
label="Guidance scale",
|
291 |
+
minimum=0.0,
|
292 |
+
maximum=10.0,
|
293 |
+
step=0.1,
|
294 |
+
value=8.0,
|
295 |
+
)
|
296 |
+
num_inference_steps = gr.Slider(
|
297 |
+
label="Number of inference steps",
|
298 |
+
minimum=1,
|
299 |
+
maximum=80,
|
300 |
+
step=1,
|
301 |
+
value=25,
|
302 |
+
)
|
303 |
+
with gr.Row():
|
304 |
+
sample_mid_interpolation = gr.Slider(
|
305 |
+
label="Number of sampling steps in the middle of interpolation",
|
306 |
+
minimum=2,
|
307 |
+
maximum=80,
|
308 |
+
step=2,
|
309 |
+
value=16,
|
310 |
+
)
|
311 |
+
num_interpolation_steps.change(
|
312 |
+
fn=update_sampling_steps,
|
313 |
+
inputs=[num_interpolation_steps, sample_mid_interpolation],
|
314 |
+
outputs=[sample_mid_interpolation],
|
315 |
+
)
|
316 |
+
with gr.Row():
|
317 |
+
remove_n_middle = gr.Slider(
|
318 |
+
label="Number of middle steps to remove from interpolation",
|
319 |
+
minimum=0,
|
320 |
+
maximum=80,
|
321 |
+
step=2,
|
322 |
+
value=0,
|
323 |
+
)
|
324 |
+
with gr.Column():
|
325 |
+
result = gr.Image(label="Result", show_label=False)
|
326 |
+
|
327 |
+
gr.Markdown(
|
328 |
+
"""
|
329 |
+
Metadata:
|
330 |
+
"""
|
331 |
+
)
|
332 |
+
with gr.Row():
|
333 |
+
show_seed = gr.Label(label="Seed:", value="Randomized seed")
|
334 |
+
ssim_score = gr.Label(
|
335 |
+
label="SSIM Score:", value="Generate to see score"
|
336 |
+
)
|
337 |
+
cos_sim = gr.Label(label="CLIP Score:", value="Generate to see score")
|
338 |
+
gr.Markdown(
|
339 |
+
f"""
|
340 |
+
Currently running on {power_device}.
|
341 |
+
"""
|
342 |
+
)
|
343 |
+
|
344 |
+
run_button.click(
|
345 |
+
fn=infer,
|
346 |
+
inputs=[
|
347 |
+
input_image,
|
348 |
+
prompt1,
|
349 |
+
prompt2,
|
350 |
+
negative_prompt,
|
351 |
+
seed,
|
352 |
+
randomize_seed,
|
353 |
+
width,
|
354 |
+
height,
|
355 |
+
guidance_scale,
|
356 |
+
interpolation_step,
|
357 |
+
num_inference_steps,
|
358 |
+
num_interpolation_steps,
|
359 |
+
sample_mid_interpolation,
|
360 |
+
remove_n_middle,
|
361 |
+
],
|
362 |
+
outputs=[result, show_seed, ssim_score, cos_sim],
|
363 |
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
364 |
|
365 |
+
demo.queue().launch(show_error=True)
|
366 |
|
367 |
"""
|
368 |
input_image,
|