Spaces:
Build error
Build error
Delete stable_cascade.py
Browse files- stable_cascade.py +0 -153
stable_cascade.py
DELETED
@@ -1,153 +0,0 @@
|
|
1 |
-
import torch, os
|
2 |
-
from diffusers import StableCascadeDecoderPipeline, StableCascadePriorPipeline
|
3 |
-
import gradio as gr
|
4 |
-
|
5 |
-
prior = StableCascadePriorPipeline.from_pretrained("stabilityai/stable-cascade-prior", torch_dtype=torch.bfloat16).to("cuda")
|
6 |
-
decoder = StableCascadeDecoderPipeline.from_pretrained("stabilityai/stable-cascade", torch_dtype=torch.float16).to("cuda")
|
7 |
-
|
8 |
-
def generate_images(
|
9 |
-
prompt="a photo of a girl",
|
10 |
-
negative_prompt="bad,ugly,deformed",
|
11 |
-
height=1024,
|
12 |
-
width=1024,
|
13 |
-
guidance_scale=4.0,
|
14 |
-
seed=42,
|
15 |
-
num_images_per_prompt=1,
|
16 |
-
prior_inference_steps=20,
|
17 |
-
decoder_inference_steps=10
|
18 |
-
):
|
19 |
-
"""
|
20 |
-
Generates images based on a given prompt using Stable Diffusion models on CUDA device.
|
21 |
-
Parameters:
|
22 |
-
- prompt (str): The prompt to generate images for.
|
23 |
-
- negative_prompt (str): The negative prompt to guide image generation away from.
|
24 |
-
- height (int): The height of the generated images.
|
25 |
-
- width (int): The width of the generated images.
|
26 |
-
- guidance_scale (float): The scale of guidance for the image generation.
|
27 |
-
- prior_inference_steps (int): The number of inference steps for the prior model.
|
28 |
-
- decoder_inference_steps (int): The number of inference steps for the decoder model.
|
29 |
-
Returns:
|
30 |
-
- List[PIL.Image]: A list of generated PIL Image objects.
|
31 |
-
"""
|
32 |
-
generator = torch.Generator(device="cuda").manual_seed(int(seed))
|
33 |
-
|
34 |
-
# Generate image embeddings using the prior model
|
35 |
-
prior_output = prior(
|
36 |
-
prompt=prompt,
|
37 |
-
generator=generator,
|
38 |
-
height=height,
|
39 |
-
width=width,
|
40 |
-
negative_prompt=negative_prompt,
|
41 |
-
guidance_scale=guidance_scale,
|
42 |
-
num_images_per_prompt=num_images_per_prompt,
|
43 |
-
num_inference_steps=prior_inference_steps
|
44 |
-
)
|
45 |
-
|
46 |
-
# Generate images using the decoder model and the embeddings from the prior model
|
47 |
-
decoder_output = decoder(
|
48 |
-
image_embeddings=prior_output.image_embeddings.half(),
|
49 |
-
prompt=prompt,
|
50 |
-
generator=generator,
|
51 |
-
negative_prompt=negative_prompt,
|
52 |
-
guidance_scale=0.0, # Guidance scale typically set to 0 for decoder as guidance is applied in the prior
|
53 |
-
output_type="pil",
|
54 |
-
num_inference_steps=decoder_inference_steps
|
55 |
-
).images
|
56 |
-
|
57 |
-
return decoder_output
|
58 |
-
|
59 |
-
|
60 |
-
def web_demo():
|
61 |
-
with gr.Blocks():
|
62 |
-
with gr.Row():
|
63 |
-
with gr.Column():
|
64 |
-
text2image_prompt = gr.Textbox(
|
65 |
-
lines=1,
|
66 |
-
placeholder="Prompt",
|
67 |
-
show_label=False,
|
68 |
-
)
|
69 |
-
|
70 |
-
text2image_negative_prompt = gr.Textbox(
|
71 |
-
lines=1,
|
72 |
-
placeholder="Negative Prompt",
|
73 |
-
show_label=False,
|
74 |
-
)
|
75 |
-
|
76 |
-
text2image_seed = gr.Number(
|
77 |
-
value=42,
|
78 |
-
label="Seed",
|
79 |
-
)
|
80 |
-
|
81 |
-
with gr.Row():
|
82 |
-
with gr.Column():
|
83 |
-
text2image_num_images_per_prompt = gr.Slider(
|
84 |
-
minimum=1,
|
85 |
-
maximum=2,
|
86 |
-
step=1,
|
87 |
-
value=1,
|
88 |
-
label="Number Image",
|
89 |
-
)
|
90 |
-
|
91 |
-
text2image_height = gr.Slider(
|
92 |
-
minimum=128,
|
93 |
-
maximum=1024,
|
94 |
-
step=32,
|
95 |
-
value=1024,
|
96 |
-
label="Image Height",
|
97 |
-
)
|
98 |
-
|
99 |
-
text2image_width = gr.Slider(
|
100 |
-
minimum=128,
|
101 |
-
maximum=1024,
|
102 |
-
step=32,
|
103 |
-
value=1024,
|
104 |
-
label="Image Width",
|
105 |
-
)
|
106 |
-
with gr.Row():
|
107 |
-
with gr.Column():
|
108 |
-
text2image_guidance_scale = gr.Slider(
|
109 |
-
minimum=0.1,
|
110 |
-
maximum=15,
|
111 |
-
step=0.1,
|
112 |
-
value=4.0,
|
113 |
-
label="Guidance Scale",
|
114 |
-
)
|
115 |
-
text2image_prior_inference_step = gr.Slider(
|
116 |
-
minimum=1,
|
117 |
-
maximum=50,
|
118 |
-
step=1,
|
119 |
-
value=20,
|
120 |
-
label="Prior Inference Step",
|
121 |
-
)
|
122 |
-
|
123 |
-
text2image_decoder_inference_step = gr.Slider(
|
124 |
-
minimum=1,
|
125 |
-
maximum=50,
|
126 |
-
step=1,
|
127 |
-
value=10,
|
128 |
-
label="Decoder Inference Step",
|
129 |
-
)
|
130 |
-
text2image_predict = gr.Button(value="Generate Image")
|
131 |
-
|
132 |
-
with gr.Column():
|
133 |
-
output_image = gr.Gallery(
|
134 |
-
label="Generated images",
|
135 |
-
show_label=False,
|
136 |
-
elem_id="gallery",
|
137 |
-
).style(grid=(1, 2), height=300)
|
138 |
-
|
139 |
-
text2image_predict.click(
|
140 |
-
fn=generate_images,
|
141 |
-
inputs=[
|
142 |
-
text2image_prompt,
|
143 |
-
text2image_negative_prompt,
|
144 |
-
text2image_height,
|
145 |
-
text2image_width,
|
146 |
-
text2image_guidance_scale,
|
147 |
-
text2image_seed,
|
148 |
-
text2image_num_images_per_prompt,
|
149 |
-
text2image_prior_inference_step,
|
150 |
-
text2image_decoder_inference_step
|
151 |
-
],
|
152 |
-
outputs=output_image,
|
153 |
-
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|