Spaces:
Running
on
A10G
Running
on
A10G
darkstorm2150
commited on
Commit
•
ba2b82b
1
Parent(s):
0e6b2f0
Update app.py
Browse filesRestoring until fix is found later
app.py
CHANGED
@@ -22,8 +22,18 @@ class Model:
|
|
22 |
def __init__(self, name, path=""):
|
23 |
self.name = name
|
24 |
self.path = path
|
25 |
-
|
26 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
27 |
|
28 |
|
29 |
models = [
|
@@ -41,19 +51,6 @@ MODELS = {m.name: m for m in models}
|
|
41 |
|
42 |
device = "GPU 🔥" if torch.cuda.is_available() else "CPU 🥶"
|
43 |
|
44 |
-
def get_model(name):
|
45 |
-
model = MODELS[name]
|
46 |
-
|
47 |
-
if model.pipe_t2i is None:
|
48 |
-
model.pipe_t2i = StableDiffusionPipeline.from_pretrained(
|
49 |
-
model.path, torch_dtype=torch.float16, safety_checker=SAFETY_CHECKER
|
50 |
-
)
|
51 |
-
model.pipe_t2i.scheduler = DPMSolverMultistepScheduler.from_config(
|
52 |
-
model.pipe_t2i.scheduler.config
|
53 |
-
)
|
54 |
-
model.pipe_i2i = StableDiffusionImg2ImgPipeline(**model.pipe_t2i.components)
|
55 |
-
|
56 |
-
return model
|
57 |
|
58 |
def error_str(error, title="Error"):
|
59 |
return (
|
@@ -63,6 +60,7 @@ def error_str(error, title="Error"):
|
|
63 |
else ""
|
64 |
)
|
65 |
|
|
|
66 |
def inference(
|
67 |
model_name,
|
68 |
prompt,
|
@@ -137,12 +135,9 @@ def txt_to_img(
|
|
137 |
):
|
138 |
pipe = MODELS[model_name].pipe_t2i
|
139 |
|
140 |
-
if
|
141 |
-
|
142 |
-
|
143 |
-
pipe.enable_xformers_memory_efficient_attention()
|
144 |
-
else:
|
145 |
-
raise ValueError(f"Unable to find pipeline for model: {model_name}")
|
146 |
|
147 |
result = pipe(
|
148 |
prompt,
|
@@ -155,12 +150,12 @@ def txt_to_img(
|
|
155 |
generator=generator,
|
156 |
)
|
157 |
|
158 |
-
|
159 |
-
|
160 |
-
torch.cuda.empty_cache()
|
161 |
|
162 |
return replace_nsfw_images(result)
|
163 |
|
|
|
164 |
def img_to_img(
|
165 |
model_name,
|
166 |
prompt,
|
@@ -175,14 +170,11 @@ def img_to_img(
|
|
175 |
generator,
|
176 |
seed,
|
177 |
):
|
178 |
-
pipe =
|
179 |
|
180 |
-
if
|
181 |
-
|
182 |
-
|
183 |
-
pipe.enable_xformers_memory_efficient_attention()
|
184 |
-
else:
|
185 |
-
raise ValueError(f"Unable to find pipeline for model: {model_name}")
|
186 |
|
187 |
ratio = min(height / img.height, width / img.width)
|
188 |
img = img.resize((int(img.width * ratio), int(img.height * ratio)), Image.LANCZOS)
|
@@ -198,18 +190,19 @@ def img_to_img(
|
|
198 |
generator=generator,
|
199 |
)
|
200 |
|
201 |
-
|
202 |
-
|
203 |
-
torch.cuda.empty_cache()
|
204 |
|
205 |
return replace_nsfw_images(result)
|
206 |
|
|
|
207 |
def replace_nsfw_images(results):
|
208 |
for i in range(len(results.images)):
|
209 |
if results.nsfw_content_detected[i]:
|
210 |
results.images[i] = Image.open("nsfw.png")
|
211 |
return results.images
|
212 |
|
|
|
213 |
with gr.Blocks(css="style.css") as demo:
|
214 |
gr.HTML(
|
215 |
f"""
|
|
|
22 |
def __init__(self, name, path=""):
|
23 |
self.name = name
|
24 |
self.path = path
|
25 |
+
|
26 |
+
if path != "":
|
27 |
+
self.pipe_t2i = StableDiffusionPipeline.from_pretrained(
|
28 |
+
path, torch_dtype=torch.float16, safety_checker=SAFETY_CHECKER
|
29 |
+
)
|
30 |
+
self.pipe_t2i.scheduler = DPMSolverMultistepScheduler.from_config(
|
31 |
+
self.pipe_t2i.scheduler.config
|
32 |
+
)
|
33 |
+
self.pipe_i2i = StableDiffusionImg2ImgPipeline(**self.pipe_t2i.components)
|
34 |
+
else:
|
35 |
+
self.pipe_t2i = None
|
36 |
+
self.pipe_i2i = None
|
37 |
|
38 |
|
39 |
models = [
|
|
|
51 |
|
52 |
device = "GPU 🔥" if torch.cuda.is_available() else "CPU 🥶"
|
53 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
54 |
|
55 |
def error_str(error, title="Error"):
|
56 |
return (
|
|
|
60 |
else ""
|
61 |
)
|
62 |
|
63 |
+
|
64 |
def inference(
|
65 |
model_name,
|
66 |
prompt,
|
|
|
135 |
):
|
136 |
pipe = MODELS[model_name].pipe_t2i
|
137 |
|
138 |
+
if torch.cuda.is_available():
|
139 |
+
pipe = pipe.to("cuda")
|
140 |
+
pipe.enable_xformers_memory_efficient_attention()
|
|
|
|
|
|
|
141 |
|
142 |
result = pipe(
|
143 |
prompt,
|
|
|
150 |
generator=generator,
|
151 |
)
|
152 |
|
153 |
+
pipe.to("cpu")
|
154 |
+
torch.cuda.empty_cache()
|
|
|
155 |
|
156 |
return replace_nsfw_images(result)
|
157 |
|
158 |
+
|
159 |
def img_to_img(
|
160 |
model_name,
|
161 |
prompt,
|
|
|
170 |
generator,
|
171 |
seed,
|
172 |
):
|
173 |
+
pipe = MODELS[model_name].pipe_i2i
|
174 |
|
175 |
+
if torch.cuda.is_available():
|
176 |
+
pipe = pipe.to("cuda")
|
177 |
+
pipe.enable_xformers_memory_efficient_attention()
|
|
|
|
|
|
|
178 |
|
179 |
ratio = min(height / img.height, width / img.width)
|
180 |
img = img.resize((int(img.width * ratio), int(img.height * ratio)), Image.LANCZOS)
|
|
|
190 |
generator=generator,
|
191 |
)
|
192 |
|
193 |
+
pipe.to("cpu")
|
194 |
+
torch.cuda.empty_cache()
|
|
|
195 |
|
196 |
return replace_nsfw_images(result)
|
197 |
|
198 |
+
|
199 |
def replace_nsfw_images(results):
|
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 |
+
|
206 |
with gr.Blocks(css="style.css") as demo:
|
207 |
gr.HTML(
|
208 |
f"""
|