add net choices
Browse files
app.py
CHANGED
@@ -1,16 +1,23 @@
|
|
1 |
import gradio as gr
|
2 |
import torch
|
3 |
from carvekit.api.interface import Interface
|
|
|
|
|
4 |
from carvekit.ml.wrap.fba_matting import FBAMatting
|
5 |
from carvekit.ml.wrap.tracer_b7 import TracerUniversalB7
|
|
|
6 |
from carvekit.pipelines.postprocessing import MattingMethod
|
7 |
from carvekit.pipelines.preprocessing import PreprocessingStub
|
8 |
from carvekit.trimap.generator import TrimapGenerator
|
9 |
|
10 |
device = 'cuda' if torch.cuda.is_available() else 'cpu'
|
11 |
|
12 |
-
|
13 |
-
|
|
|
|
|
|
|
|
|
14 |
|
15 |
fba = FBAMatting(device=device,
|
16 |
input_tensor_size=2048,
|
@@ -24,18 +31,19 @@ postprocessing = MattingMethod(matting_module=fba,
|
|
24 |
trimap_generator=trimap,
|
25 |
device=device)
|
26 |
|
27 |
-
|
28 |
-
post_pipe=postprocessing,
|
29 |
-
seg_pipe=seg_net)
|
30 |
|
31 |
|
32 |
-
def generate_trimap(original):
|
33 |
-
mask =
|
34 |
return trimap(original_image=original, mask=mask[0])
|
35 |
|
36 |
|
37 |
-
def predict(image):
|
38 |
-
|
|
|
|
|
|
|
39 |
|
40 |
|
41 |
footer = r"""
|
@@ -49,35 +57,37 @@ Demo based on <a href='https://github.com/OPHoperHPO/image-background-remove-too
|
|
49 |
"""
|
50 |
|
51 |
with gr.Blocks(title="CarveKit") as app:
|
52 |
-
|
53 |
gr.Markdown("<center><h1><b>CarveKit</b></h1></center>")
|
54 |
gr.HTML("<center><h3>High-quality image background removal</h3></center>")
|
55 |
|
56 |
with gr.Tabs() as tabs:
|
57 |
with gr.TabItem("Remove background", id=0):
|
58 |
-
with gr.Row(
|
59 |
with gr.Column():
|
60 |
input_img = gr.Image(type="pil", label="Input image")
|
|
|
|
|
|
|
|
|
61 |
run_btn = gr.Button(variant="primary")
|
62 |
with gr.Column():
|
63 |
output_img = gr.Image(type="pil", label="result")
|
64 |
|
65 |
-
run_btn.click(predict, [input_img], [output_img])
|
66 |
|
67 |
with gr.TabItem("Trimap generator", id=1):
|
68 |
-
with gr.Row(
|
69 |
with gr.Column():
|
70 |
trimap_input = gr.Image(type="pil", label="Input image")
|
|
|
|
|
|
|
|
|
71 |
trimap_btn = gr.Button(variant="primary")
|
72 |
with gr.Column():
|
73 |
trimap_output = gr.Image(type="pil", label="result")
|
74 |
|
75 |
-
trimap_btn.click(generate_trimap, [trimap_input], [trimap_output])
|
76 |
-
|
77 |
-
# with gr.Row():
|
78 |
-
# examples_data = [[f"examples/{x:02d}.jpg"] for x in range(1, 4)]
|
79 |
-
# examples = gr.Dataset(components=[input_img], samples=examples_data)
|
80 |
-
# examples.click(lambda x: x[0], [examples], [input_img])
|
81 |
|
82 |
with gr.Row():
|
83 |
gr.HTML(footer)
|
|
|
1 |
import gradio as gr
|
2 |
import torch
|
3 |
from carvekit.api.interface import Interface
|
4 |
+
from carvekit.ml.wrap.basnet import BASNET
|
5 |
+
from carvekit.ml.wrap.deeplab_v3 import DeepLabV3
|
6 |
from carvekit.ml.wrap.fba_matting import FBAMatting
|
7 |
from carvekit.ml.wrap.tracer_b7 import TracerUniversalB7
|
8 |
+
from carvekit.ml.wrap.u2net import U2NET
|
9 |
from carvekit.pipelines.postprocessing import MattingMethod
|
10 |
from carvekit.pipelines.preprocessing import PreprocessingStub
|
11 |
from carvekit.trimap.generator import TrimapGenerator
|
12 |
|
13 |
device = 'cuda' if torch.cuda.is_available() else 'cpu'
|
14 |
|
15 |
+
segment_net = {
|
16 |
+
"U2NET": U2NET(device=device, batch_size=1),
|
17 |
+
"BASNET": BASNET(device=device, batch_size=1),
|
18 |
+
"DeepLabV3": DeepLabV3(device=device, batch_size=1),
|
19 |
+
"TracerUniversalB7": TracerUniversalB7(device=device, batch_size=1)
|
20 |
+
}
|
21 |
|
22 |
fba = FBAMatting(device=device,
|
23 |
input_tensor_size=2048,
|
|
|
31 |
trimap_generator=trimap,
|
32 |
device=device)
|
33 |
|
34 |
+
method_choices = [k for k, v in segment_net.items()]
|
|
|
|
|
35 |
|
36 |
|
37 |
+
def generate_trimap(method, original):
|
38 |
+
mask = segment_net[method]([original])
|
39 |
return trimap(original_image=original, mask=mask[0])
|
40 |
|
41 |
|
42 |
+
def predict(method, image):
|
43 |
+
method = segment_net[method]
|
44 |
+
return Interface(pre_pipe=preprocessing,
|
45 |
+
post_pipe=postprocessing,
|
46 |
+
seg_pipe=method)([image])[0]
|
47 |
|
48 |
|
49 |
footer = r"""
|
|
|
57 |
"""
|
58 |
|
59 |
with gr.Blocks(title="CarveKit") as app:
|
|
|
60 |
gr.Markdown("<center><h1><b>CarveKit</b></h1></center>")
|
61 |
gr.HTML("<center><h3>High-quality image background removal</h3></center>")
|
62 |
|
63 |
with gr.Tabs() as tabs:
|
64 |
with gr.TabItem("Remove background", id=0):
|
65 |
+
with gr.Row(equal_height=False):
|
66 |
with gr.Column():
|
67 |
input_img = gr.Image(type="pil", label="Input image")
|
68 |
+
drp_itf = gr.Dropdown(
|
69 |
+
value="TracerUniversalB7",
|
70 |
+
label="Segmentor model",
|
71 |
+
choices=method_choices)
|
72 |
run_btn = gr.Button(variant="primary")
|
73 |
with gr.Column():
|
74 |
output_img = gr.Image(type="pil", label="result")
|
75 |
|
76 |
+
run_btn.click(predict, [drp_itf, input_img], [output_img])
|
77 |
|
78 |
with gr.TabItem("Trimap generator", id=1):
|
79 |
+
with gr.Row(equal_height=False):
|
80 |
with gr.Column():
|
81 |
trimap_input = gr.Image(type="pil", label="Input image")
|
82 |
+
drp_itf = gr.Dropdown(
|
83 |
+
value="TracerUniversalB7",
|
84 |
+
label="Segmentor model",
|
85 |
+
choices=method_choices)
|
86 |
trimap_btn = gr.Button(variant="primary")
|
87 |
with gr.Column():
|
88 |
trimap_output = gr.Image(type="pil", label="result")
|
89 |
|
90 |
+
trimap_btn.click(generate_trimap, [drp_itf, trimap_input], [trimap_output])
|
|
|
|
|
|
|
|
|
|
|
91 |
|
92 |
with gr.Row():
|
93 |
gr.HTML(footer)
|