Spaces:
Runtime error
Runtime error
Updates
Browse files- app.py +20 -15
- requirements.txt +2 -4
app.py
CHANGED
@@ -20,11 +20,11 @@ if not torch.cuda.is_available():
|
|
20 |
DESCRIPTION += "\n<p>Running on CPU 🥶</p>"
|
21 |
|
22 |
MAX_SEED = np.iinfo(np.int32).max
|
23 |
-
CACHE_EXAMPLES = torch.cuda.is_available() and os.getenv("CACHE_EXAMPLES") != "0"
|
24 |
MAX_IMAGE_SIZE = int(os.getenv("MAX_IMAGE_SIZE", "1536"))
|
25 |
USE_TORCH_COMPILE = False
|
26 |
ENABLE_CPU_OFFLOAD = os.getenv("ENABLE_CPU_OFFLOAD") == "1"
|
27 |
-
PREVIEW_IMAGES =
|
28 |
|
29 |
dtype = torch.bfloat16
|
30 |
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
|
@@ -47,10 +47,12 @@ if torch.cuda.is_available():
|
|
47 |
previewer = Previewer()
|
48 |
previewer_state_dict = torch.load("previewer/previewer_v1_100k.pt", map_location=torch.device('cpu'))["state_dict"]
|
49 |
previewer.load_state_dict(previewer_state_dict)
|
50 |
-
def callback_prior(
|
|
|
51 |
output = previewer(latents)
|
52 |
output = numpy_to_pil(output.clamp(0, 1).permute(0, 2, 3, 1).float().cpu().numpy())
|
53 |
-
|
|
|
54 |
callback_steps = 1
|
55 |
else:
|
56 |
previewer = None
|
@@ -62,6 +64,7 @@ else:
|
|
62 |
|
63 |
|
64 |
def randomize_seed_fn(seed: int, randomize_seed: bool) -> int:
|
|
|
65 |
if randomize_seed:
|
66 |
seed = random.randint(0, MAX_SEED)
|
67 |
return seed
|
@@ -82,7 +85,8 @@ def generate(
|
|
82 |
num_images_per_prompt: int = 2,
|
83 |
profile: gr.OAuthProfile | None = None,
|
84 |
) -> PIL.Image.Image:
|
85 |
-
|
|
|
86 |
prior_pipeline.to(device)
|
87 |
decoder_pipeline.to(device)
|
88 |
|
@@ -98,10 +102,9 @@ def generate(
|
|
98 |
guidance_scale=prior_guidance_scale,
|
99 |
num_images_per_prompt=num_images_per_prompt,
|
100 |
generator=generator,
|
101 |
-
|
102 |
-
|
103 |
)
|
104 |
-
|
105 |
if PREVIEW_IMAGES:
|
106 |
for _ in range(len(DEFAULT_STAGE_C_TIMESTEPS)):
|
107 |
r = next(prior_output)
|
@@ -119,7 +122,7 @@ def generate(
|
|
119 |
generator=generator,
|
120 |
output_type="pil",
|
121 |
).images
|
122 |
-
|
123 |
#Save images
|
124 |
for image in decoder_output:
|
125 |
user_history.save_image(
|
@@ -137,15 +140,17 @@ def generate(
|
|
137 |
"num_images_per_prompt": num_images_per_prompt,
|
138 |
},
|
139 |
)
|
140 |
-
|
141 |
yield decoder_output[0]
|
142 |
|
143 |
|
144 |
examples = [
|
145 |
-
"
|
146 |
-
"
|
147 |
-
"
|
148 |
-
"
|
|
|
|
|
149 |
]
|
150 |
|
151 |
with gr.Blocks() as demo:
|
@@ -277,4 +282,4 @@ with gr.Blocks(css="style.css") as demo_with_history:
|
|
277 |
user_history.render()
|
278 |
|
279 |
if __name__ == "__main__":
|
280 |
-
demo_with_history.queue(max_size=20).launch()
|
|
|
20 |
DESCRIPTION += "\n<p>Running on CPU 🥶</p>"
|
21 |
|
22 |
MAX_SEED = np.iinfo(np.int32).max
|
23 |
+
CACHE_EXAMPLES = False #torch.cuda.is_available() and os.getenv("CACHE_EXAMPLES") != "0"
|
24 |
MAX_IMAGE_SIZE = int(os.getenv("MAX_IMAGE_SIZE", "1536"))
|
25 |
USE_TORCH_COMPILE = False
|
26 |
ENABLE_CPU_OFFLOAD = os.getenv("ENABLE_CPU_OFFLOAD") == "1"
|
27 |
+
PREVIEW_IMAGES = False
|
28 |
|
29 |
dtype = torch.bfloat16
|
30 |
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
|
|
|
47 |
previewer = Previewer()
|
48 |
previewer_state_dict = torch.load("previewer/previewer_v1_100k.pt", map_location=torch.device('cpu'))["state_dict"]
|
49 |
previewer.load_state_dict(previewer_state_dict)
|
50 |
+
def callback_prior(pipeline, step_index, t, callback_kwargs):
|
51 |
+
latents = callback_kwargs["latents"]
|
52 |
output = previewer(latents)
|
53 |
output = numpy_to_pil(output.clamp(0, 1).permute(0, 2, 3, 1).float().cpu().numpy())
|
54 |
+
callback_kwargs["preview_output"] = output
|
55 |
+
return callback_kwargs
|
56 |
callback_steps = 1
|
57 |
else:
|
58 |
previewer = None
|
|
|
64 |
|
65 |
|
66 |
def randomize_seed_fn(seed: int, randomize_seed: bool) -> int:
|
67 |
+
print("randomizing seed")
|
68 |
if randomize_seed:
|
69 |
seed = random.randint(0, MAX_SEED)
|
70 |
return seed
|
|
|
85 |
num_images_per_prompt: int = 2,
|
86 |
profile: gr.OAuthProfile | None = None,
|
87 |
) -> PIL.Image.Image:
|
88 |
+
|
89 |
+
#previewer.eval().requires_grad_(False).to(device).to(dtype)
|
90 |
prior_pipeline.to(device)
|
91 |
decoder_pipeline.to(device)
|
92 |
|
|
|
102 |
guidance_scale=prior_guidance_scale,
|
103 |
num_images_per_prompt=num_images_per_prompt,
|
104 |
generator=generator,
|
105 |
+
#callback_on_step_end=callback_prior,
|
106 |
+
#callback_on_step_end_tensor_inputs=['latents']
|
107 |
)
|
|
|
108 |
if PREVIEW_IMAGES:
|
109 |
for _ in range(len(DEFAULT_STAGE_C_TIMESTEPS)):
|
110 |
r = next(prior_output)
|
|
|
122 |
generator=generator,
|
123 |
output_type="pil",
|
124 |
).images
|
125 |
+
print(decoder_output)
|
126 |
#Save images
|
127 |
for image in decoder_output:
|
128 |
user_history.save_image(
|
|
|
140 |
"num_images_per_prompt": num_images_per_prompt,
|
141 |
},
|
142 |
)
|
143 |
+
|
144 |
yield decoder_output[0]
|
145 |
|
146 |
|
147 |
examples = [
|
148 |
+
"A futuristic cityscape at sunset",
|
149 |
+
"pair of shoes made of dried fruit skins, 3d render, bright colours, clean composition, beautiful artwork, logo",
|
150 |
+
"post-apocalyptic wasteland, the most delicate beautiful flower with green leaves growing from dust and rubble, vibrant colours, cinematic",
|
151 |
+
"Mixed media artwork, Emotional cyborg girl, Elegant dress, Skin lesions as a storytelling element, In the style of surrealist expressionism, muted color scheme, dreamlike atmosphere, abstract and distorted forms",
|
152 |
+
"rendering, side shot, falf-strange body with complex system equipment with hyper detail robot, gaze, sci-fi, gloomy environment, foggy with light shader, cyan and yellow illuminations, dramatic lighting, RTX shader, hyper detail texture with reflection, HDRI, cyborg, grunge, bolt, UHD",
|
153 |
+
"vintage Japanese postcard, in the style of Kentaro Miura, featuring a black cat holding a vinyl record in its paws, with vintage colors including light beige and red tones on a white background, very detailed artwork."
|
154 |
]
|
155 |
|
156 |
with gr.Blocks() as demo:
|
|
|
282 |
user_history.render()
|
283 |
|
284 |
if __name__ == "__main__":
|
285 |
+
demo_with_history.queue(max_size=20).launch()
|
requirements.txt
CHANGED
@@ -1,5 +1,3 @@
|
|
1 |
-
|
2 |
-
https://gradio-builds.s3.amazonaws.com/aabb08191a7d94d2a1e9ff87b0d3c3987cd519c5/gradio-4.18.0-py3-none-any.whl
|
3 |
accelerate
|
4 |
-
|
5 |
-
transformers
|
|
|
1 |
+
diffusers
|
|
|
2 |
accelerate
|
3 |
+
transformers
|
|