Spaces:
Runtime error
Runtime error
scheduler for sd v2
Browse files
app.py
CHANGED
@@ -37,19 +37,6 @@ models = [
|
|
37 |
Model("Pony Diffusion", "AstraliteHeart/pony-diffusion"),
|
38 |
Model("Robo Diffusion", "nousr/robo-diffusion"),
|
39 |
]
|
40 |
-
|
41 |
-
scheduler = DPMSolverMultistepScheduler(
|
42 |
-
beta_start=0.00085,
|
43 |
-
beta_end=0.012,
|
44 |
-
beta_schedule="scaled_linear",
|
45 |
-
num_train_timesteps=1000,
|
46 |
-
trained_betas=None,
|
47 |
-
predict_epsilon=True,
|
48 |
-
thresholding=False,
|
49 |
-
algorithm_type="dpmsolver++",
|
50 |
-
solver_type="midpoint",
|
51 |
-
lower_order_final=True,
|
52 |
-
)
|
53 |
|
54 |
custom_model = None
|
55 |
if is_colab:
|
@@ -61,23 +48,20 @@ current_model = models[1] if is_colab else models[0]
|
|
61 |
current_model_path = current_model.path
|
62 |
|
63 |
if is_colab:
|
64 |
-
pipe = StableDiffusionPipeline.from_pretrained(
|
65 |
-
|
66 |
-
|
67 |
-
|
68 |
-
|
69 |
-
|
70 |
-
|
71 |
-
|
72 |
-
|
73 |
-
|
74 |
-
|
75 |
-
|
76 |
-
|
77 |
-
|
78 |
-
# models.remove(model)
|
79 |
-
# pipe = models[0].pipe_t2i
|
80 |
-
|
81 |
if torch.cuda.is_available():
|
82 |
pipe = pipe.to("cuda")
|
83 |
|
@@ -98,7 +82,7 @@ def on_model_change(model_name):
|
|
98 |
|
99 |
return gr.update(visible = model_name == models[0].name), gr.update(placeholder=prefix)
|
100 |
|
101 |
-
def inference(model_name, prompt, guidance, steps, width=512, height=512, seed=0, img=None, strength=0.5, neg_prompt=""):
|
102 |
|
103 |
print(psutil.virtual_memory()) # print memory usage
|
104 |
|
@@ -112,13 +96,13 @@ def inference(model_name, prompt, guidance, steps, width=512, height=512, seed=0
|
|
112 |
|
113 |
try:
|
114 |
if img is not None:
|
115 |
-
return img_to_img(model_path, prompt, neg_prompt, img, strength, guidance, steps, width, height, generator), None
|
116 |
else:
|
117 |
-
return txt_to_img(model_path, prompt, neg_prompt, guidance, steps, width, height, generator), None
|
118 |
except Exception as e:
|
119 |
return None, error_str(e)
|
120 |
|
121 |
-
def txt_to_img(model_path, prompt, neg_prompt, guidance, steps, width, height, generator):
|
122 |
|
123 |
print(f"{datetime.datetime.now()} txt_to_img, model: {current_model.name}")
|
124 |
|
@@ -129,9 +113,18 @@ def txt_to_img(model_path, prompt, neg_prompt, guidance, steps, width, height, g
|
|
129 |
current_model_path = model_path
|
130 |
|
131 |
if is_colab or current_model == custom_model:
|
132 |
-
pipe = StableDiffusionPipeline.from_pretrained(
|
|
|
|
|
|
|
|
|
|
|
133 |
else:
|
134 |
-
pipe = StableDiffusionPipeline.from_pretrained(
|
|
|
|
|
|
|
|
|
135 |
# pipe = pipe.to("cpu")
|
136 |
# pipe = current_model.pipe_t2i
|
137 |
|
@@ -143,7 +136,7 @@ def txt_to_img(model_path, prompt, neg_prompt, guidance, steps, width, height, g
|
|
143 |
result = pipe(
|
144 |
prompt,
|
145 |
negative_prompt = neg_prompt,
|
146 |
-
|
147 |
num_inference_steps = int(steps),
|
148 |
guidance_scale = guidance,
|
149 |
width = width,
|
@@ -152,7 +145,7 @@ def txt_to_img(model_path, prompt, neg_prompt, guidance, steps, width, height, g
|
|
152 |
|
153 |
return replace_nsfw_images(result)
|
154 |
|
155 |
-
def img_to_img(model_path, prompt, neg_prompt, img, strength, guidance, steps, width, height, generator):
|
156 |
|
157 |
print(f"{datetime.datetime.now()} img_to_img, model: {model_path}")
|
158 |
|
@@ -163,9 +156,18 @@ def img_to_img(model_path, prompt, neg_prompt, img, strength, guidance, steps, w
|
|
163 |
current_model_path = model_path
|
164 |
|
165 |
if is_colab or current_model == custom_model:
|
166 |
-
pipe = StableDiffusionImg2ImgPipeline.from_pretrained(
|
|
|
|
|
|
|
|
|
|
|
167 |
else:
|
168 |
-
pipe = StableDiffusionImg2ImgPipeline.from_pretrained(
|
|
|
|
|
|
|
|
|
169 |
# pipe = pipe.to("cpu")
|
170 |
# pipe = current_model.pipe_i2i
|
171 |
|
@@ -179,7 +181,7 @@ def img_to_img(model_path, prompt, neg_prompt, img, strength, guidance, steps, w
|
|
179 |
result = pipe(
|
180 |
prompt,
|
181 |
negative_prompt = neg_prompt,
|
182 |
-
|
183 |
init_image = img,
|
184 |
num_inference_steps = int(steps),
|
185 |
strength = strength,
|
@@ -193,12 +195,12 @@ def img_to_img(model_path, prompt, neg_prompt, img, strength, guidance, steps, w
|
|
193 |
def replace_nsfw_images(results):
|
194 |
|
195 |
if is_colab:
|
196 |
-
return results.images
|
197 |
|
198 |
for i in range(len(results.images)):
|
199 |
if results.nsfw_content_detected[i]:
|
200 |
results.images[i] = Image.open("nsfw.png")
|
201 |
-
return results.images
|
202 |
|
203 |
css = """.finetuned-diffusion-div div{display:inline-flex;align-items:center;gap:.8rem;font-size:1.75rem}.finetuned-diffusion-div div h1{font-weight:900;margin-bottom:7px}.finetuned-diffusion-div p{margin-bottom:10px;font-size:94%}a{text-decoration:underline}.tabs{margin-top:0;margin-bottom:0}#gallery{min-height:20rem}
|
204 |
"""
|
@@ -216,7 +218,8 @@ with gr.Blocks(css=css) as demo:
|
|
216 |
<p>You can skip the queue and load custom models in the colab: <a href="https://colab.research.google.com/gist/qunash/42112fb104509c24fd3aa6d1c11dd6e0/copy-of-fine-tuned-diffusion-gradio.ipynb"><img data-canonical-src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab" src="https://camo.githubusercontent.com/84f0493939e0c4de4e6dbe113251b4bfb5353e57134ffd9fcab6b8714514d4d1/68747470733a2f2f636f6c61622e72657365617263682e676f6f676c652e636f6d2f6173736574732f636f6c61622d62616467652e737667"></a></p>
|
217 |
Running on <b>{device}</b>{(" in a <b>Google Colab</b>." if is_colab else "")}
|
218 |
</p>
|
219 |
-
<p>You can also duplicate this space and upgrade to gpu by going to settings
|
|
|
220 |
</div>
|
221 |
"""
|
222 |
)
|
@@ -234,10 +237,9 @@ with gr.Blocks(css=css) as demo:
|
|
234 |
generate = gr.Button(value="Generate").style(rounded=(False, True, True, False))
|
235 |
|
236 |
|
237 |
-
image_out = gr.Image(height=512)
|
238 |
-
|
239 |
-
|
240 |
-
# ).style(grid=[1], height="auto")
|
241 |
error_output = gr.Markdown()
|
242 |
|
243 |
with gr.Column(scale=45):
|
@@ -245,7 +247,7 @@ with gr.Blocks(css=css) as demo:
|
|
245 |
with gr.Group():
|
246 |
neg_prompt = gr.Textbox(label="Negative prompt", placeholder="What to exclude from the image")
|
247 |
|
248 |
-
|
249 |
|
250 |
with gr.Row():
|
251 |
guidance = gr.Slider(label="Guidance scale", value=7.5, maximum=15)
|
@@ -267,18 +269,18 @@ with gr.Blocks(css=css) as demo:
|
|
267 |
custom_model_path.change(custom_model_changed, inputs=custom_model_path, outputs=None)
|
268 |
# n_images.change(lambda n: gr.Gallery().style(grid=[2 if n > 1 else 1], height="auto"), inputs=n_images, outputs=gallery)
|
269 |
|
270 |
-
inputs = [model_name, prompt, guidance, steps, width, height, seed, image, strength, neg_prompt]
|
271 |
-
outputs = [
|
272 |
prompt.submit(inference, inputs=inputs, outputs=outputs)
|
273 |
generate.click(inference, inputs=inputs, outputs=outputs)
|
274 |
|
275 |
ex = gr.Examples([
|
276 |
-
[models[7].name, "tiny cute and adorable kitten adventurer dressed in a warm overcoat with survival gear on a winters day", 7.5,
|
277 |
-
[models[4].name, "portrait of dwayne johnson", 7.0,
|
278 |
-
[models[5].name, "portrait of a beautiful alyx vance half life", 10,
|
279 |
-
[models[6].name, "Aloy from Horizon: Zero Dawn, half body portrait, smooth, detailed armor, beautiful face, illustration", 7.0,
|
280 |
-
[models[5].name, "fantasy portrait painting, digital art", 4.0,
|
281 |
-
], inputs=[model_name, prompt, guidance, steps
|
282 |
|
283 |
gr.HTML("""
|
284 |
<div style="border-top: 1px solid #303030;">
|
|
|
37 |
Model("Pony Diffusion", "AstraliteHeart/pony-diffusion"),
|
38 |
Model("Robo Diffusion", "nousr/robo-diffusion"),
|
39 |
]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
40 |
|
41 |
custom_model = None
|
42 |
if is_colab:
|
|
|
48 |
current_model_path = current_model.path
|
49 |
|
50 |
if is_colab:
|
51 |
+
pipe = StableDiffusionPipeline.from_pretrained(
|
52 |
+
current_model.path,
|
53 |
+
torch_dtype=torch.float16,
|
54 |
+
scheduler=DPMSolverMultistepScheduler.from_pretrained(current_model.path, subfolder="scheduler"),
|
55 |
+
safety_checker=lambda images, clip_input: (images, False)
|
56 |
+
)
|
57 |
+
|
58 |
+
else:
|
59 |
+
pipe = StableDiffusionPipeline.from_pretrained(
|
60 |
+
current_model.path,
|
61 |
+
torch_dtype=torch.float16,
|
62 |
+
scheduler=DPMSolverMultistepScheduler.from_pretrained(current_model.path, subfolder="scheduler")
|
63 |
+
)
|
64 |
+
|
|
|
|
|
|
|
65 |
if torch.cuda.is_available():
|
66 |
pipe = pipe.to("cuda")
|
67 |
|
|
|
82 |
|
83 |
return gr.update(visible = model_name == models[0].name), gr.update(placeholder=prefix)
|
84 |
|
85 |
+
def inference(model_name, prompt, guidance, steps, n_images=1, width=512, height=512, seed=0, img=None, strength=0.5, neg_prompt=""):
|
86 |
|
87 |
print(psutil.virtual_memory()) # print memory usage
|
88 |
|
|
|
96 |
|
97 |
try:
|
98 |
if img is not None:
|
99 |
+
return img_to_img(model_path, prompt, n_images, neg_prompt, img, strength, guidance, steps, width, height, generator), None
|
100 |
else:
|
101 |
+
return txt_to_img(model_path, prompt, n_images, neg_prompt, guidance, steps, width, height, generator), None
|
102 |
except Exception as e:
|
103 |
return None, error_str(e)
|
104 |
|
105 |
+
def txt_to_img(model_path, prompt, n_images, neg_prompt, guidance, steps, width, height, generator):
|
106 |
|
107 |
print(f"{datetime.datetime.now()} txt_to_img, model: {current_model.name}")
|
108 |
|
|
|
113 |
current_model_path = model_path
|
114 |
|
115 |
if is_colab or current_model == custom_model:
|
116 |
+
pipe = StableDiffusionPipeline.from_pretrained(
|
117 |
+
current_model_path,
|
118 |
+
torch_dtype=torch.float16,
|
119 |
+
scheduler=DPMSolverMultistepScheduler.from_pretrained(current_model.path, subfolder="scheduler"),
|
120 |
+
safety_checker=lambda images, clip_input: (images, False)
|
121 |
+
)
|
122 |
else:
|
123 |
+
pipe = StableDiffusionPipeline.from_pretrained(
|
124 |
+
current_model_path,
|
125 |
+
torch_dtype=torch.float16,
|
126 |
+
scheduler=DPMSolverMultistepScheduler.from_pretrained(current_model.path, subfolder="scheduler")
|
127 |
+
)
|
128 |
# pipe = pipe.to("cpu")
|
129 |
# pipe = current_model.pipe_t2i
|
130 |
|
|
|
136 |
result = pipe(
|
137 |
prompt,
|
138 |
negative_prompt = neg_prompt,
|
139 |
+
num_images_per_prompt=n_images,
|
140 |
num_inference_steps = int(steps),
|
141 |
guidance_scale = guidance,
|
142 |
width = width,
|
|
|
145 |
|
146 |
return replace_nsfw_images(result)
|
147 |
|
148 |
+
def img_to_img(model_path, prompt, n_images, neg_prompt, img, strength, guidance, steps, width, height, generator):
|
149 |
|
150 |
print(f"{datetime.datetime.now()} img_to_img, model: {model_path}")
|
151 |
|
|
|
156 |
current_model_path = model_path
|
157 |
|
158 |
if is_colab or current_model == custom_model:
|
159 |
+
pipe = StableDiffusionImg2ImgPipeline.from_pretrained(
|
160 |
+
current_model_path,
|
161 |
+
torch_dtype=torch.float16,
|
162 |
+
scheduler=DPMSolverMultistepScheduler.from_pretrained(current_model.path, subfolder="scheduler"),
|
163 |
+
safety_checker=lambda images, clip_input: (images, False)
|
164 |
+
)
|
165 |
else:
|
166 |
+
pipe = StableDiffusionImg2ImgPipeline.from_pretrained(
|
167 |
+
current_model_path,
|
168 |
+
torch_dtype=torch.float16,
|
169 |
+
scheduler=DPMSolverMultistepScheduler.from_pretrained(current_model.path, subfolder="scheduler")
|
170 |
+
)
|
171 |
# pipe = pipe.to("cpu")
|
172 |
# pipe = current_model.pipe_i2i
|
173 |
|
|
|
181 |
result = pipe(
|
182 |
prompt,
|
183 |
negative_prompt = neg_prompt,
|
184 |
+
num_images_per_prompt=n_images,
|
185 |
init_image = img,
|
186 |
num_inference_steps = int(steps),
|
187 |
strength = strength,
|
|
|
195 |
def replace_nsfw_images(results):
|
196 |
|
197 |
if is_colab:
|
198 |
+
return results.images
|
199 |
|
200 |
for i in range(len(results.images)):
|
201 |
if results.nsfw_content_detected[i]:
|
202 |
results.images[i] = Image.open("nsfw.png")
|
203 |
+
return results.images
|
204 |
|
205 |
css = """.finetuned-diffusion-div div{display:inline-flex;align-items:center;gap:.8rem;font-size:1.75rem}.finetuned-diffusion-div div h1{font-weight:900;margin-bottom:7px}.finetuned-diffusion-div p{margin-bottom:10px;font-size:94%}a{text-decoration:underline}.tabs{margin-top:0;margin-bottom:0}#gallery{min-height:20rem}
|
206 |
"""
|
|
|
218 |
<p>You can skip the queue and load custom models in the colab: <a href="https://colab.research.google.com/gist/qunash/42112fb104509c24fd3aa6d1c11dd6e0/copy-of-fine-tuned-diffusion-gradio.ipynb"><img data-canonical-src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab" src="https://camo.githubusercontent.com/84f0493939e0c4de4e6dbe113251b4bfb5353e57134ffd9fcab6b8714514d4d1/68747470733a2f2f636f6c61622e72657365617263682e676f6f676c652e636f6d2f6173736574732f636f6c61622d62616467652e737667"></a></p>
|
219 |
Running on <b>{device}</b>{(" in a <b>Google Colab</b>." if is_colab else "")}
|
220 |
</p>
|
221 |
+
<p>You can also duplicate this space and upgrade to gpu by going to settings:<br>
|
222 |
+
<a style="display:inline-block" href="https://huggingface.co/spaces/anzorq/finetuned_diffusion?duplicate=true"><img src="https://img.shields.io/badge/-Duplicate%20Space-blue?labelColor=white&style=flat&logo=data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAABAAAAAQCAYAAAAf8/9hAAAAAXNSR0IArs4c6QAAAP5JREFUOE+lk7FqAkEURY+ltunEgFXS2sZGIbXfEPdLlnxJyDdYB62sbbUKpLbVNhyYFzbrrA74YJlh9r079973psed0cvUD4A+4HoCjsA85X0Dfn/RBLBgBDxnQPfAEJgBY+A9gALA4tcbamSzS4xq4FOQAJgCDwV2CPKV8tZAJcAjMMkUe1vX+U+SMhfAJEHasQIWmXNN3abzDwHUrgcRGmYcgKe0bxrblHEB4E/pndMazNpSZGcsZdBlYJcEL9Afo75molJyM2FxmPgmgPqlWNLGfwZGG6UiyEvLzHYDmoPkDDiNm9JR9uboiONcBXrpY1qmgs21x1QwyZcpvxt9NS09PlsPAAAAAElFTkSuQmCC&logoWidth=14" alt="Duplicate Space"></a></p>
|
223 |
</div>
|
224 |
"""
|
225 |
)
|
|
|
237 |
generate = gr.Button(value="Generate").style(rounded=(False, True, True, False))
|
238 |
|
239 |
|
240 |
+
# image_out = gr.Image(height=512)
|
241 |
+
gallery = gr.Gallery(label="Generated images", show_label=False, elem_id="gallery").style(grid=[2], height="auto")
|
242 |
+
|
|
|
243 |
error_output = gr.Markdown()
|
244 |
|
245 |
with gr.Column(scale=45):
|
|
|
247 |
with gr.Group():
|
248 |
neg_prompt = gr.Textbox(label="Negative prompt", placeholder="What to exclude from the image")
|
249 |
|
250 |
+
n_images = gr.Slider(label="Images", value=1, minimum=1, maximum=4, step=1)
|
251 |
|
252 |
with gr.Row():
|
253 |
guidance = gr.Slider(label="Guidance scale", value=7.5, maximum=15)
|
|
|
269 |
custom_model_path.change(custom_model_changed, inputs=custom_model_path, outputs=None)
|
270 |
# n_images.change(lambda n: gr.Gallery().style(grid=[2 if n > 1 else 1], height="auto"), inputs=n_images, outputs=gallery)
|
271 |
|
272 |
+
inputs = [model_name, prompt, guidance, steps, n_images, width, height, seed, image, strength, neg_prompt]
|
273 |
+
outputs = [gallery, error_output]
|
274 |
prompt.submit(inference, inputs=inputs, outputs=outputs)
|
275 |
generate.click(inference, inputs=inputs, outputs=outputs)
|
276 |
|
277 |
ex = gr.Examples([
|
278 |
+
[models[7].name, "tiny cute and adorable kitten adventurer dressed in a warm overcoat with survival gear on a winters day", 7.5, 25],
|
279 |
+
[models[4].name, "portrait of dwayne johnson", 7.0, 35],
|
280 |
+
[models[5].name, "portrait of a beautiful alyx vance half life", 10, 25],
|
281 |
+
[models[6].name, "Aloy from Horizon: Zero Dawn, half body portrait, smooth, detailed armor, beautiful face, illustration", 7.0, 30],
|
282 |
+
[models[5].name, "fantasy portrait painting, digital art", 4.0, 20],
|
283 |
+
], inputs=[model_name, prompt, guidance, steps], outputs=outputs, fn=inference, cache_examples=False)
|
284 |
|
285 |
gr.HTML("""
|
286 |
<div style="border-top: 1px solid #303030;">
|