Spaces:
Running
on
Zero
Running
on
Zero
Update app.py
Browse files
app.py
CHANGED
@@ -5,58 +5,12 @@ import diffusers
|
|
5 |
import os
|
6 |
from PIL import Image
|
7 |
hf_token = os.environ.get("HF_TOKEN")
|
8 |
-
from diffusers import
|
9 |
-
|
10 |
-
ratios_map = {
|
11 |
-
0.5:{"width":704,"height":1408},
|
12 |
-
0.57:{"width":768,"height":1344},
|
13 |
-
0.68:{"width":832,"height":1216},
|
14 |
-
0.72:{"width":832,"height":1152},
|
15 |
-
0.78:{"width":896,"height":1152},
|
16 |
-
0.82:{"width":896,"height":1088},
|
17 |
-
0.88:{"width":960,"height":1088},
|
18 |
-
0.94:{"width":960,"height":1024},
|
19 |
-
1.00:{"width":1024,"height":1024},
|
20 |
-
1.13:{"width":1088,"height":960},
|
21 |
-
1.21:{"width":1088,"height":896},
|
22 |
-
1.29:{"width":1152,"height":896},
|
23 |
-
1.38:{"width":1152,"height":832},
|
24 |
-
1.46:{"width":1216,"height":832},
|
25 |
-
1.67:{"width":1280,"height":768},
|
26 |
-
1.75:{"width":1344,"height":768},
|
27 |
-
2.00:{"width":1408,"height":704}
|
28 |
-
}
|
29 |
-
ratios = np.array(list(ratios_map.keys()))
|
30 |
-
|
31 |
-
def get_size(init_image):
|
32 |
-
w,h=init_image.size
|
33 |
-
curr_ratio = w/h
|
34 |
-
ind = np.argmin(np.abs(curr_ratio-ratios))
|
35 |
-
ratio = ratios[ind]
|
36 |
-
chosen_ratio = ratios_map[ratio]
|
37 |
-
w,h = chosen_ratio['width'], chosen_ratio['height']
|
38 |
|
39 |
-
return w,h
|
40 |
|
41 |
device = "cuda" if torch.cuda.is_available() else "cpu"
|
42 |
-
|
43 |
-
|
44 |
-
subfolder="unet",
|
45 |
-
torch_dtype=torch.float16,
|
46 |
-
)
|
47 |
-
|
48 |
-
scheduler = DDIMScheduler.from_pretrained("briaai/BRIA-2.3", subfolder="scheduler",
|
49 |
-
rescale_betas_zero_snr=True,prediction_type='v_prediction',timestep_spacing="trailing",clip_sample=False)
|
50 |
-
|
51 |
-
pipe = StableDiffusionXLInpaintPipeline.from_pretrained(
|
52 |
-
"briaai/BRIA-2.3",
|
53 |
-
unet=unet,
|
54 |
-
scheduler=scheduler,
|
55 |
-
torch_dtype=torch.float16,
|
56 |
-
force_zeros_for_empty_prompt=False
|
57 |
-
)
|
58 |
-
pipe = pipe.to(device)
|
59 |
-
pipe.force_zeros_for_empty_prompt = False
|
60 |
|
61 |
default_negative_prompt= "" #"Logo,Watermark,Text,Ugly,Morbid,Extra fingers,Poorly drawn hands,Mutation,Blurry,Extra limbs,Gross proportions,Missing arms,Mutated hands,Long neck,Duplicate,Mutilated,Mutilated hands,Poorly drawn face,Deformed,Bad anatomy,Cloned face,Malformed limbs,Missing legs,Too many fingers"
|
62 |
|
@@ -69,27 +23,28 @@ def read_content(file_path: str) -> str:
|
|
69 |
|
70 |
return content
|
71 |
|
72 |
-
def predict(dict, prompt="", negative_prompt="", guidance_scale=5, steps=30,
|
73 |
-
if negative_prompt == "":
|
74 |
-
negative_prompt = None
|
75 |
|
76 |
-
|
77 |
-
init_image = dict["image"].convert("RGB")#.resize((1024, 1024))
|
78 |
-
mask = dict["mask"].convert("RGB")#.resize((1024, 1024))
|
79 |
-
|
80 |
-
w,h = get_size(init_image)
|
81 |
|
82 |
-
|
83 |
-
|
84 |
-
|
85 |
-
# Resize to nearest ratio ?
|
86 |
-
|
87 |
-
mask = np.array(mask)
|
88 |
-
mask[mask>0]=255
|
89 |
-
mask = Image.fromarray(mask)
|
90 |
|
91 |
-
|
|
|
92 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
93 |
return output.images[0] #, gr.update(visible=True)
|
94 |
|
95 |
|
@@ -146,14 +101,20 @@ with image_blocks as demo:
|
|
146 |
image = gr.Image(sources=['upload'], tool='sketch', elem_id="image_upload", type="pil", label="Upload", height=400)
|
147 |
with gr.Row(elem_id="prompt-container", equal_height=True):
|
148 |
with gr.Row():
|
149 |
-
prompt = gr.Textbox(placeholder="Your prompt (
|
150 |
-
btn = gr.Button("
|
|
|
|
|
|
|
|
|
|
|
|
|
151 |
|
152 |
with gr.Accordion(label="Advanced Settings", open=False):
|
153 |
with gr.Row(equal_height=True):
|
154 |
guidance_scale = gr.Number(value=5, minimum=1.0, maximum=10.0, step=0.5, label="guidance_scale")
|
155 |
-
steps = gr.Number(value=30, minimum=
|
156 |
-
|
157 |
negative_prompt = gr.Textbox(label="negative_prompt", value=default_negative_prompt, placeholder=default_negative_prompt, info="what you don't want to see in the image")
|
158 |
|
159 |
|
@@ -162,8 +123,8 @@ with image_blocks as demo:
|
|
162 |
|
163 |
|
164 |
|
165 |
-
btn.click(fn=predict, inputs=[image, prompt, negative_prompt, guidance_scale, steps,
|
166 |
-
prompt.submit(fn=predict, inputs=[image, prompt, negative_prompt, guidance_scale, steps,
|
167 |
|
168 |
# gr.Examples(
|
169 |
# examples=[
|
|
|
5 |
import os
|
6 |
from PIL import Image
|
7 |
hf_token = os.environ.get("HF_TOKEN")
|
8 |
+
from diffusers import AutoPipelineForText2Image
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
9 |
|
|
|
10 |
|
11 |
device = "cuda" if torch.cuda.is_available() else "cpu"
|
12 |
+
pipe = AutoPipelineForText2Image.from_pretrained("briaai/BRIA-2.3", torch_dtype=torch.float16, force_zeros_for_empty_prompt=False).to(device)
|
13 |
+
pipe.load_ip_adapter("briaai/DEV-Image-Prompt", subfolder='models', weight_name="ip_adapter_bria.bin")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
14 |
|
15 |
default_negative_prompt= "" #"Logo,Watermark,Text,Ugly,Morbid,Extra fingers,Poorly drawn hands,Mutation,Blurry,Extra limbs,Gross proportions,Missing arms,Mutated hands,Long neck,Duplicate,Mutilated,Mutilated hands,Poorly drawn face,Deformed,Bad anatomy,Cloned face,Malformed limbs,Missing legs,Too many fingers"
|
16 |
|
|
|
23 |
|
24 |
return content
|
25 |
|
26 |
+
def predict(dict, prompt="high quality, best quality", negative_prompt="", guidance_scale=5, steps=30, ip_adapter_scale = 1.0, width=1024, height=1024, seed=0):
|
|
|
|
|
27 |
|
28 |
+
pipeline.set_ip_adapter_scale(ip_adapter_scale)
|
|
|
|
|
|
|
|
|
29 |
|
30 |
+
if negative_prompt == "":
|
31 |
+
negative_prompt = None
|
|
|
|
|
|
|
|
|
|
|
|
|
32 |
|
33 |
+
init_image = dict["image"].convert("RGB")
|
34 |
+
init_image = init_image.resize((224, 224))
|
35 |
|
36 |
+
generator = torch.Generator(device="cpu").manual_seed(seed)
|
37 |
+
|
38 |
+
output = pipe(
|
39 |
+
prompt=prompt,
|
40 |
+
negative_prompt=negative_prompt,
|
41 |
+
ip_adapter_image=init_image,
|
42 |
+
num_inference_steps=steps,
|
43 |
+
generator=generator,
|
44 |
+
height=height, width=width,
|
45 |
+
guidance_scale=guidance_scale
|
46 |
+
).images
|
47 |
+
|
48 |
return output.images[0] #, gr.update(visible=True)
|
49 |
|
50 |
|
|
|
101 |
image = gr.Image(sources=['upload'], tool='sketch', elem_id="image_upload", type="pil", label="Upload", height=400)
|
102 |
with gr.Row(elem_id="prompt-container", equal_height=True):
|
103 |
with gr.Row():
|
104 |
+
prompt = gr.Textbox(placeholder="Your prompt (you can leave it empty if you only want the image prompt as input)", show_label=False, elem_id="prompt")
|
105 |
+
btn = gr.Button("Generate!", elem_id="run_button")
|
106 |
+
|
107 |
+
with gr.Accordion(label="Settings", open=True):
|
108 |
+
with gr.Row(equal_height=True):
|
109 |
+
ip_adapter_scale = gr.Number(value=1.0, minimum=0.01, maximum=1.0, step=0.01, label="ip_adapter_scale")
|
110 |
+
width = gr.Number(value=1024, minimum=0.01, maximum=1.0, step=0.01, label="width")
|
111 |
+
height = gr.Number(value=1024, minimum=0.01, maximum=1.0, step=0.01, label="height")
|
112 |
|
113 |
with gr.Accordion(label="Advanced Settings", open=False):
|
114 |
with gr.Row(equal_height=True):
|
115 |
guidance_scale = gr.Number(value=5, minimum=1.0, maximum=10.0, step=0.5, label="guidance_scale")
|
116 |
+
steps = gr.Number(value=30, minimum=10, maximum=100, step=1, label="steps")
|
117 |
+
seed = gr.Number(value=0, minimum=0, maximum=100000, step=1, label="seed")
|
118 |
negative_prompt = gr.Textbox(label="negative_prompt", value=default_negative_prompt, placeholder=default_negative_prompt, info="what you don't want to see in the image")
|
119 |
|
120 |
|
|
|
123 |
|
124 |
|
125 |
|
126 |
+
btn.click(fn=predict, inputs=[image, prompt, negative_prompt, guidance_scale, steps, ip_adapter_scale, width, height, seed], outputs=[image_out], api_name='run')
|
127 |
+
prompt.submit(fn=predict, inputs=[image, prompt, negative_prompt, guidance_scale, steps, ip_adapter_scale, width, height, seed], outputs=[image_out])
|
128 |
|
129 |
# gr.Examples(
|
130 |
# examples=[
|