Spaces:
Runtime error
Runtime error
Jordan Legg
commited on
Commit
Β·
da39f41
1
Parent(s):
f071803
shaping latents
Browse files
app.py
CHANGED
@@ -8,37 +8,18 @@ from torchvision import transforms
|
|
8 |
from diffusers import DiffusionPipeline
|
9 |
|
10 |
# Define constants
|
|
|
|
|
11 |
MAX_SEED = np.iinfo(np.int32).max
|
12 |
MAX_IMAGE_SIZE = 2048
|
13 |
-
MIN_IMAGE_SIZE = 256
|
14 |
-
DEFAULT_IMAGE_SIZE = 1024
|
15 |
-
MAX_PROMPT_LENGTH = 256 # Changed to 256 as per FLUX.1-schnell requirements
|
16 |
-
|
17 |
-
# Check for GPU availability
|
18 |
-
device = "cuda" if torch.cuda.is_available() else "cpu"
|
19 |
-
if device == "cpu":
|
20 |
-
print("Warning: Running on CPU. This may be very slow.")
|
21 |
-
|
22 |
-
dtype = torch.float16 if device == "cuda" else torch.float32
|
23 |
-
|
24 |
-
def load_model():
|
25 |
-
try:
|
26 |
-
pipe = DiffusionPipeline.from_pretrained("black-forest-labs/FLUX.1-schnell", torch_dtype=dtype)
|
27 |
-
pipe.to(device)
|
28 |
-
pipe.enable_model_cpu_offload()
|
29 |
-
pipe.vae.enable_slicing()
|
30 |
-
pipe.vae.enable_tiling()
|
31 |
-
return pipe
|
32 |
-
except Exception as e:
|
33 |
-
raise RuntimeError(f"Failed to load the model: {str(e)}")
|
34 |
|
35 |
# Load the diffusion pipeline
|
36 |
-
pipe =
|
37 |
|
38 |
-
def preprocess_image(image
|
39 |
# Preprocess the image for the VAE
|
40 |
preprocess = transforms.Compose([
|
41 |
-
transforms.Resize(
|
42 |
transforms.ToTensor(),
|
43 |
transforms.Normalize([0.5], [0.5])
|
44 |
])
|
@@ -51,76 +32,51 @@ def encode_image(image, vae):
|
|
51 |
latents = vae.encode(image).latent_dist.sample() * 0.18215
|
52 |
return latents
|
53 |
|
54 |
-
def validate_inputs(prompt, width, height, num_inference_steps):
|
55 |
-
if not prompt or len(prompt) > MAX_PROMPT_LENGTH:
|
56 |
-
raise ValueError(f"Prompt must be between 1 and {MAX_PROMPT_LENGTH} characters.")
|
57 |
-
if width % 8 != 0 or height % 8 != 0:
|
58 |
-
raise ValueError("Width and height must be divisible by 8.")
|
59 |
-
if width < MIN_IMAGE_SIZE or width > MAX_IMAGE_SIZE or height < MIN_IMAGE_SIZE or height > MAX_IMAGE_SIZE:
|
60 |
-
raise ValueError(f"Image dimensions must be between {MIN_IMAGE_SIZE} and {MAX_IMAGE_SIZE}.")
|
61 |
-
if num_inference_steps < 1 or num_inference_steps > 50:
|
62 |
-
raise ValueError("Number of inference steps must be between 1 and 50.")
|
63 |
-
|
64 |
@spaces.GPU()
|
65 |
-
def infer(prompt, init_image=None, seed=42, randomize_seed=False, width=
|
66 |
-
|
67 |
-
|
68 |
-
|
69 |
-
|
70 |
-
|
71 |
-
|
72 |
-
|
73 |
-
|
74 |
-
|
|
|
|
|
|
|
75 |
|
76 |
-
|
77 |
-
|
78 |
-
init_image = init_image.convert("RGB")
|
79 |
-
init_image = preprocess_image(init_image, (height, width))
|
80 |
-
|
81 |
-
# Encode the image using the VAE
|
82 |
-
init_latents = encode_image(init_image, pipe.vae)
|
83 |
-
|
84 |
-
# Ensure latents are correctly shaped
|
85 |
-
init_latents = torch.nn.functional.interpolate(init_latents, size=(height // 8, width // 8), mode='bilinear', align_corners=False)
|
86 |
-
|
87 |
-
# Add noise to latents
|
88 |
-
noise = torch.randn_like(init_latents)
|
89 |
-
latents = noise + strength * (init_latents - noise)
|
90 |
-
|
91 |
-
image = pipe(
|
92 |
-
prompt=prompt,
|
93 |
-
height=height,
|
94 |
-
width=width,
|
95 |
-
num_inference_steps=num_inference_steps,
|
96 |
-
generator=generator,
|
97 |
-
guidance_scale=0.0,
|
98 |
-
latents=latents,
|
99 |
-
max_sequence_length=max_sequence_length
|
100 |
-
).images[0]
|
101 |
-
else:
|
102 |
-
# Process text2img
|
103 |
-
image = pipe(
|
104 |
-
prompt=prompt,
|
105 |
-
height=height,
|
106 |
-
width=width,
|
107 |
-
num_inference_steps=num_inference_steps,
|
108 |
-
generator=generator,
|
109 |
-
guidance_scale=0.0,
|
110 |
-
max_sequence_length=max_sequence_length
|
111 |
-
).images[0]
|
112 |
|
113 |
-
|
114 |
-
|
115 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
116 |
|
117 |
# Define example prompts
|
118 |
examples = [
|
119 |
"a tiny astronaut hatching from an egg on the moon",
|
120 |
"a cat holding a sign that says hello world",
|
121 |
"an anime illustration of a wiener schnitzel",
|
122 |
-
"a surreal landscape with floating islands and waterfalls",
|
123 |
-
"a steampunk-inspired cityscape at sunset"
|
124 |
]
|
125 |
|
126 |
# CSS styling for the Japanese-inspired interface
|
@@ -173,7 +129,7 @@ with gr.Blocks(css=css) as demo:
|
|
173 |
label="Prompt",
|
174 |
show_label=False,
|
175 |
max_lines=1,
|
176 |
-
placeholder=
|
177 |
container=False,
|
178 |
)
|
179 |
run_button = gr.Button("Run", scale=0)
|
@@ -195,17 +151,17 @@ with gr.Blocks(css=css) as demo:
|
|
195 |
with gr.Row():
|
196 |
width = gr.Slider(
|
197 |
label="Width",
|
198 |
-
minimum=
|
199 |
maximum=MAX_IMAGE_SIZE,
|
200 |
-
step=
|
201 |
-
value=
|
202 |
)
|
203 |
height = gr.Slider(
|
204 |
label="Height",
|
205 |
-
minimum=
|
206 |
maximum=MAX_IMAGE_SIZE,
|
207 |
-
step=
|
208 |
-
value=
|
209 |
)
|
210 |
|
211 |
with gr.Row():
|
@@ -216,13 +172,6 @@ with gr.Blocks(css=css) as demo:
|
|
216 |
step=1,
|
217 |
value=4,
|
218 |
)
|
219 |
-
strength = gr.Slider(
|
220 |
-
label="Strength (for img2img)",
|
221 |
-
minimum=0.0,
|
222 |
-
maximum=1.0,
|
223 |
-
step=0.01,
|
224 |
-
value=0.8,
|
225 |
-
)
|
226 |
|
227 |
gr.Examples(
|
228 |
examples=examples,
|
@@ -235,9 +184,8 @@ with gr.Blocks(css=css) as demo:
|
|
235 |
gr.on(
|
236 |
triggers=[run_button.click, prompt.submit],
|
237 |
fn=infer,
|
238 |
-
inputs=[prompt, init_image, seed, randomize_seed, width, height, num_inference_steps
|
239 |
outputs=[result, seed]
|
240 |
)
|
241 |
|
242 |
-
|
243 |
-
demo.launch()
|
|
|
8 |
from diffusers import DiffusionPipeline
|
9 |
|
10 |
# Define constants
|
11 |
+
dtype = torch.bfloat16
|
12 |
+
device = "cuda" if torch.cuda.is_available() else "cpu"
|
13 |
MAX_SEED = np.iinfo(np.int32).max
|
14 |
MAX_IMAGE_SIZE = 2048
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
15 |
|
16 |
# Load the diffusion pipeline
|
17 |
+
pipe = DiffusionPipeline.from_pretrained("black-forest-labs/FLUX.1-schnell", torch_dtype=dtype).to(device)
|
18 |
|
19 |
+
def preprocess_image(image):
|
20 |
# Preprocess the image for the VAE
|
21 |
preprocess = transforms.Compose([
|
22 |
+
transforms.Resize((512, 512)), # Adjust the size as needed
|
23 |
transforms.ToTensor(),
|
24 |
transforms.Normalize([0.5], [0.5])
|
25 |
])
|
|
|
32 |
latents = vae.encode(image).latent_dist.sample() * 0.18215
|
33 |
return latents
|
34 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
35 |
@spaces.GPU()
|
36 |
+
def infer(prompt, init_image=None, seed=42, randomize_seed=False, width=1024, height=1024, num_inference_steps=4, progress=gr.Progress(track_tqdm=True)):
|
37 |
+
if randomize_seed:
|
38 |
+
seed = random.randint(0, MAX_SEED)
|
39 |
+
generator = torch.Generator().manual_seed(seed)
|
40 |
+
|
41 |
+
if init_image is not None:
|
42 |
+
# Process img2img
|
43 |
+
init_image = init_image.convert("RGB")
|
44 |
+
init_image = preprocess_image(init_image)
|
45 |
+
latents = encode_image(init_image, pipe.vae)
|
46 |
+
# Ensure latents are correctly shaped and adjusted
|
47 |
+
latents = torch.nn.functional.interpolate(latents, size=(height // 8, width // 8))
|
48 |
+
latents = latents * 0.18215 # Adjust latent scaling factor if necessary
|
49 |
|
50 |
+
# Ensure latents are reshaped to match the expected input dimensions of the model
|
51 |
+
latents = latents.view(1, -1, height // 8, width // 8)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
52 |
|
53 |
+
image = pipe(
|
54 |
+
prompt=prompt,
|
55 |
+
height=height,
|
56 |
+
width=width,
|
57 |
+
num_inference_steps=num_inference_steps,
|
58 |
+
generator=generator,
|
59 |
+
guidance_scale=0.0,
|
60 |
+
latents=latents
|
61 |
+
).images[0]
|
62 |
+
else:
|
63 |
+
# Process text2img
|
64 |
+
image = pipe(
|
65 |
+
prompt=prompt,
|
66 |
+
height=height,
|
67 |
+
width=width,
|
68 |
+
num_inference_steps=num_inference_steps,
|
69 |
+
generator=generator,
|
70 |
+
guidance_scale=0.0
|
71 |
+
).images[0]
|
72 |
+
|
73 |
+
return image, seed
|
74 |
|
75 |
# Define example prompts
|
76 |
examples = [
|
77 |
"a tiny astronaut hatching from an egg on the moon",
|
78 |
"a cat holding a sign that says hello world",
|
79 |
"an anime illustration of a wiener schnitzel",
|
|
|
|
|
80 |
]
|
81 |
|
82 |
# CSS styling for the Japanese-inspired interface
|
|
|
129 |
label="Prompt",
|
130 |
show_label=False,
|
131 |
max_lines=1,
|
132 |
+
placeholder="Enter your prompt",
|
133 |
container=False,
|
134 |
)
|
135 |
run_button = gr.Button("Run", scale=0)
|
|
|
151 |
with gr.Row():
|
152 |
width = gr.Slider(
|
153 |
label="Width",
|
154 |
+
minimum=256,
|
155 |
maximum=MAX_IMAGE_SIZE,
|
156 |
+
step=32,
|
157 |
+
value=1024,
|
158 |
)
|
159 |
height = gr.Slider(
|
160 |
label="Height",
|
161 |
+
minimum=256,
|
162 |
maximum=MAX_IMAGE_SIZE,
|
163 |
+
step=32,
|
164 |
+
value=1024,
|
165 |
)
|
166 |
|
167 |
with gr.Row():
|
|
|
172 |
step=1,
|
173 |
value=4,
|
174 |
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
175 |
|
176 |
gr.Examples(
|
177 |
examples=examples,
|
|
|
184 |
gr.on(
|
185 |
triggers=[run_button.click, prompt.submit],
|
186 |
fn=infer,
|
187 |
+
inputs=[prompt, init_image, seed, randomize_seed, width, height, num_inference_steps],
|
188 |
outputs=[result, seed]
|
189 |
)
|
190 |
|
191 |
+
demo.launch()
|
|