Update app.py
Browse files
app.py
CHANGED
@@ -5,45 +5,32 @@ import plotly.graph_objects as go
|
|
5 |
from sam2point import dataset
|
6 |
import sam2point.configs as configs
|
7 |
from demo_utils import run_demo, create_box
|
8 |
-
|
|
|
9 |
samples = {
|
10 |
"3D Indoor Scene - S3DIS": ["Conference Room", "Restroom", "Lobby", "Office1", "Office2"],
|
11 |
-
# "3D Indoor Scene - ScanNet": ["Scene1", "Scene2", "Scene3", "Scene4", "Scene5"],
|
12 |
"3D Indoor Scene - ScanNet": ["Scene1", "Scene2", "Scene3", "Scene4", "Scene5", "Scene6"],
|
13 |
"3D Outdoor Driving Scene - KITTI": ["Scene1", "Scene2", "Scene3", "Scene4", "Scene5", "Scene6"],
|
14 |
"3D Outdoor Street Scene - Semantic3D": ["Scene1", "Scene2", "Scene3", "Scene4", "Scene5", "Scene6", "Scene7"],
|
15 |
"3D Object - Objaverse": ["Plant", "Lego", "Lock", "Eleplant", "Knife Rest", "Skateboard", "Popcorn Machine", "Stove", "Bus Shelter", "Thor Hammer", "Horse"],
|
16 |
-
# "3D Object - Objaverse": ["Plant", "Eleplant", "Knife Rest", "Skateboard", "Popcorn Machine", "Stove", "Bus Shelter", "Thor Hammer", "Horse", "Dinner Booth"],
|
17 |
}
|
18 |
|
19 |
-
|
20 |
PATH = {
|
21 |
"S3DIS": ['Area_1_conferenceRoom_1.txt', 'Area_2_WC_1.txt', 'Area_4_lobby_2.txt', 'Area_5_office_3.txt', 'Area_6_office_9.txt'],
|
22 |
-
# "ScanNet": ['scene0001_01.pth', 'scene0005_01.pth', 'scene0010_01.pth', 'scene0016_02.pth', 'scene0019_01.pth'],
|
23 |
"ScanNet": ['scene0005_01.pth', 'scene0010_01.pth', 'scene0016_02.pth', 'scene0019_01.pth', 'scene0000_00.pth', 'scene0002_00.pth'],
|
24 |
"Objaverse": ["plant.npy", "human.npy", "lock.npy", "elephant.npy", "knife_rest.npy", "skateboard.npy", "popcorn_machine.npy", "stove.npy", "bus_shelter.npy", "thor_hammer.npy", "horse.npy"],
|
25 |
-
# "Objaverse": ["plant.npy", "elephant.npy", "knife_rest.npy", "skateboard.npy", "popcorn_machine.npy", "stove.npy", "bus_shelter.npy", "thor_hammer.npy", "horse.npy", "dinner_booth.npy"],
|
26 |
"KITTI": ["scene1.npy", "scene2.npy", "scene3.npy", "scene4.npy", "scene5.npy", "scene6.npy"],
|
27 |
"Semantic3D": ["scene1.npy", "scene2.npy", "patch19.npy", "patch0.npy", "patch1.npy", "patch50.npy", "patch62.npy"]
|
28 |
}
|
29 |
|
30 |
-
|
31 |
prompt_types = ["Point", "Box", "Mask"]
|
32 |
|
33 |
-
|
34 |
-
# def select(name, sample_idx):
|
35 |
-
# DATASET = name.split('-')[1].replace(" ", "")
|
36 |
-
# gr.Info(f"Visualizing {DATASET} Example {str(sample_idx)}...")
|
37 |
-
|
38 |
-
|
39 |
-
|
40 |
-
|
41 |
-
# Function to load and display 3D scene or object
|
42 |
def load_3d_scene(name, sample_idx=-1, type_=None, prompt=None, final=False, new_color=None):
|
43 |
DATASET = name.split('-')[1].replace(" ", "")
|
44 |
path = 'data/' + DATASET + '/' + PATH[DATASET][sample_idx]
|
45 |
asp, SIZE = 1., 1
|
46 |
-
|
47 |
print(path)
|
48 |
if DATASET == 'S3DIS':
|
49 |
point, color = dataset.load_S3DIS_sample(path, sample=True)
|
@@ -62,25 +49,14 @@ def load_3d_scene(name, sample_idx=-1, type_=None, prompt=None, final=False, new
|
|
62 |
elif DATASET == 'Semantic3D':
|
63 |
point, color = dataset.load_Semantic3D_sample(path, sample_idx, sample=True)
|
64 |
alpha = 0.2
|
65 |
-
print("Loading Dataset:", DATASET, "
|
66 |
|
67 |
-
|
68 |
-
##### Initial Showing #####
|
69 |
if not type_:
|
70 |
-
if point.shape[0] > 100000:
|
71 |
indices = np.random.choice(point.shape[0], 100000, replace=False)
|
72 |
point = point[indices]
|
73 |
color = color[indices]
|
74 |
-
# #NOTE KITTI
|
75 |
-
# mask1 = point[:, 1] <= 0.8
|
76 |
-
# mask4 = point[:, 1] >= 0.6
|
77 |
-
# mask2 = point[:, 0] >= 0.3
|
78 |
-
# mask3 = point[:, 0] <= 0.7
|
79 |
-
# mask = mask1 & mask2 & mask3 & mask4
|
80 |
-
# point = point[mask]
|
81 |
-
# color = color[mask]
|
82 |
-
# alpha = 1
|
83 |
-
# ######
|
84 |
fig = go.Figure(
|
85 |
data=[
|
86 |
go.Scatter3d(
|
@@ -101,7 +77,7 @@ def load_3d_scene(name, sample_idx=-1, type_=None, prompt=None, final=False, new
|
|
101 |
)
|
102 |
)
|
103 |
return fig
|
104 |
-
##### Final
|
105 |
if final:
|
106 |
color = new_color
|
107 |
green = np.array([[0.1, 0.1, 0.1]])
|
@@ -116,10 +92,6 @@ def load_3d_scene(name, sample_idx=-1, type_=None, prompt=None, final=False, new
|
|
116 |
indices = np.random.choice(point.shape[0], 100000, replace=False)
|
117 |
point = point[indices]
|
118 |
color = color[indices]
|
119 |
-
# mask = point[:, 1] < 0.8
|
120 |
-
# point = point[mask]
|
121 |
-
# color = color[mask]
|
122 |
-
# alpha = 1
|
123 |
scatter = go.Scatter3d(
|
124 |
x=point[:,0], y=point[:,1], z=point[:,2],
|
125 |
mode='markers',
|
@@ -128,36 +100,18 @@ def load_3d_scene(name, sample_idx=-1, type_=None, prompt=None, final=False, new
|
|
128 |
)
|
129 |
if final: scatter = [scatter, add_green] + create_box(prompt)
|
130 |
else: scatter = [scatter] + create_box(prompt)
|
131 |
-
|
132 |
elif type_ == "point":
|
133 |
prompt = np.array([prompt])
|
134 |
new = go.Scatter3d(
|
135 |
x=prompt[:,0], y=prompt[:,1], z=prompt[:,2],
|
136 |
mode='markers',
|
137 |
-
|
138 |
-
# marker=dict(size=5, color='rgb(255, 140, 0)', opacity=1),
|
139 |
-
marker=dict(size=5, color='rgb(139, 0, 0)', opacity=1),
|
140 |
name="Point Prompt"
|
141 |
)
|
142 |
-
# print(point.shape, color.shape, new_color.shape)
|
143 |
if point.shape[0] > 100000:
|
144 |
indices = np.random.choice(point.shape[0], 100000, replace=False)
|
145 |
point = point[indices]
|
146 |
color = color[indices]
|
147 |
-
# #NOTE KITTI
|
148 |
-
# mask1 = point[:, 1] <= 0.8
|
149 |
-
# mask = point[:, 1] >= 0.35 #2
|
150 |
-
# < 0.63 #3
|
151 |
-
# mask2 = point[:, 0] >= 0.3
|
152 |
-
# mask3 = point[:, 0] <= 0.7
|
153 |
-
# mask = mask1 & mask2 & mask3 & mask4
|
154 |
-
# #NOTE S3DIS
|
155 |
-
# if DATASET == 'S3DIS':
|
156 |
-
# mask = point[:, 0] > 0.04
|
157 |
-
# point = point[mask]
|
158 |
-
# color = color[mask]
|
159 |
-
# alpha = 1
|
160 |
-
# ######
|
161 |
scatter = go.Scatter3d(
|
162 |
x=point[:,0], y=point[:,1], z=point[:,2],
|
163 |
mode='markers',
|
@@ -191,12 +145,6 @@ def load_3d_scene(name, sample_idx=-1, type_=None, prompt=None, final=False, new
|
|
191 |
indices = np.random.choice(point.shape[0], 100000, replace=False)
|
192 |
point = point[indices]
|
193 |
color = color[indices]
|
194 |
-
# # cut
|
195 |
-
# mask = point[:, 0] > 0.1
|
196 |
-
# point = point[mask]
|
197 |
-
# color = color[mask]
|
198 |
-
# alpha = 1
|
199 |
-
# ######
|
200 |
scatter = go.Scatter3d(
|
201 |
x=point[:,0], y=point[:,1], z=point[:,2],
|
202 |
mode='markers',
|
@@ -204,12 +152,10 @@ def load_3d_scene(name, sample_idx=-1, type_=None, prompt=None, final=False, new
|
|
204 |
name="3D Object/Scene"
|
205 |
)
|
206 |
scatter = [scatter, add_green]
|
207 |
-
print(point.shape, color.shape)
|
208 |
else:
|
209 |
print("Wrong Prompt Type")
|
210 |
exit(1)
|
211 |
|
212 |
-
|
213 |
fig = go.Figure(
|
214 |
data=scatter,
|
215 |
layout=dict(
|
@@ -224,25 +170,21 @@ def load_3d_scene(name, sample_idx=-1, type_=None, prompt=None, final=False, new
|
|
224 |
)
|
225 |
return fig
|
226 |
|
227 |
-
|
228 |
-
|
229 |
-
|
230 |
-
# Function to display prompt in 3D
|
231 |
def show_prompt_in_3d(name, sample_idx, prompt_type, prompt_idx):
|
|
|
|
|
|
|
232 |
DATASET = name.split('-')[1].replace(" ", "")
|
233 |
TYPE = prompt_type.lower()
|
234 |
theta = 0. if DATASET in "S3DIS ScanNet" else 0.5
|
235 |
mode = "bilinear" if DATASET in "S3DIS ScanNet" else 'nearest'
|
236 |
-
|
237 |
-
|
238 |
prompt = run_demo(DATASET, TYPE, sample_idx, prompt_idx, 0.02, theta, mode, ret_prompt=True)
|
239 |
fig = load_3d_scene(name, sample_idx, TYPE, prompt)
|
240 |
-
return fig
|
241 |
-
|
242 |
|
243 |
-
|
244 |
-
|
245 |
-
# Function to start segmentation
|
246 |
def start_segmentation(name=None, sample_idx=None, prompt_type=None, prompt_idx=None, vx=0.02):
|
247 |
if name == None or sample_idx == None or prompt_type == None or prompt_idx == None:
|
248 |
return gr.Plot(), gr.Textbox(label="Response", value="Please ensure all options are selected.", visible=True)
|
@@ -252,26 +194,23 @@ def start_segmentation(name=None, sample_idx=None, prompt_type=None, prompt_idx=
|
|
252 |
theta = 0. if DATASET in "S3DIS ScanNet" else 0.5
|
253 |
mode = "bilinear" if DATASET in "S3DIS ScanNet" else 'nearest'
|
254 |
|
255 |
-
|
256 |
new_color, prompt = run_demo(DATASET, TYPE, sample_idx, prompt_idx, vx, theta, mode, ret_prompt=False)
|
257 |
fig = load_3d_scene(name, sample_idx, TYPE, prompt, final=True, new_color=new_color)
|
258 |
return fig, gr.Textbox(label="Response", value="Segmentation completed successfully!", visible=True)
|
259 |
|
260 |
-
|
261 |
-
|
262 |
-
|
263 |
def update1(datasets):
|
264 |
if 'Objaverse' in datasets:
|
265 |
-
return gr.Radio(label="Select 3D Object", choices=samples[datasets]), gr.Textbox(label="Response", value="", visible=True)
|
266 |
-
return gr.Radio(label="Select 3D Scene", choices=samples[datasets]), gr.Textbox(label="Response", value="", visible=True)
|
267 |
-
|
268 |
|
|
|
269 |
def update2(name, sample_idx, prompt_type):
|
270 |
if name == None or sample_idx == None or prompt_type == None:
|
271 |
-
return gr.Radio(label="Select Prompt Example", choices=[]), gr.Textbox(label="Response", value="", visible=True)
|
272 |
DATASET = name.split('-')[1].replace(" ", "")
|
273 |
TYPE = prompt_type.lower() + '_prompts'
|
274 |
-
|
275 |
if DATASET == 'S3DIS':
|
276 |
info = configs.S3DIS_samples[sample_idx][TYPE]
|
277 |
elif DATASET == 'ScanNet':
|
@@ -284,14 +223,15 @@ def update2(name, sample_idx, prompt_type):
|
|
284 |
info = configs.Semantic3D_samples[sample_idx][TYPE]
|
285 |
|
286 |
cur = ['Example ' + str(i) for i in range(1, len(info) + 1)]
|
287 |
-
return gr.Radio(label="Select Prompt Example", choices=cur), gr.Textbox(label="Response", value="", visible=True)
|
288 |
-
|
289 |
-
|
290 |
def update3(name, sample_idx, prompt_type, prompt_idx):
|
291 |
if name == None or sample_idx == None or prompt_type == None:
|
292 |
return gr.Textbox(label="Response", value="", visible=True), gr.Slider(minimum=0.01, maximum=0.15, step=0.001, label="Voxel Size", value=0.02)
|
293 |
DATASET = name.split('-')[1].replace(" ", "")
|
294 |
TYPE = configs.VOXEL[prompt_type.lower()]
|
|
|
295 |
if DATASET in "S3DIS ScanNet":
|
296 |
vx_ = 0.02
|
297 |
elif DATASET == 'Objaverse':
|
@@ -303,14 +243,9 @@ def update3(name, sample_idx, prompt_type, prompt_idx):
|
|
303 |
|
304 |
return gr.Textbox(label="Response", value="", visible=True), gr.Slider(minimum=0.01, maximum=0.15, step=0.001, label="Voxel Size", value=vx_)
|
305 |
|
306 |
-
|
307 |
def main():
|
308 |
-
title = """<h1 style="
|
309 |
-
<h3 align="center"><b>Segment Any 3D as Videos in Zero-shot and Promptable Manners</h3>
|
310 |
-
<br>
|
311 |
-
"""
|
312 |
-
title = """
|
313 |
-
<h1 style="text-align: center;">
|
314 |
<div style="width: 1.2em; height: 1.2em; display: inline-block;"><img src="https://github.com/ZiyuGuo99/ZiyuGuo99.github.io/blob/main/assets/img/logo.png?raw=true" style='width: 100%; height: 100%; object-fit: contain;' /></div>
|
315 |
<span style="font-variant: small-caps; font-weight: bold;">Sam2Point</span>
|
316 |
</h1>
|
@@ -351,43 +286,23 @@ def main():
|
|
351 |
prompt_type_dropdown = gr.Radio(label="Select Prompt Type", choices=prompt_types)
|
352 |
prompt_sample_dropdown = gr.Radio(label="Select Prompt Example", choices=[], type="index")
|
353 |
show_prompt_button = gr.Button("Show Prompt in 3D Scene/Object")
|
354 |
-
# show_button.input(select, [sample_dropdown, scene_dropdown], [])
|
355 |
with gr.Column():
|
356 |
-
# vx = gr.Slider(minimum=0.01, maximum=0.15, step=0.001, label="Voxel Size", value=0.02)
|
357 |
start_segment_button = gr.Button("Start Segmentation")
|
358 |
plot1 = gr.Plot()
|
359 |
|
360 |
-
|
361 |
-
|
362 |
-
|
363 |
response = gr.Textbox(label="Response")
|
364 |
|
365 |
sample_dropdown.change(update1, sample_dropdown, [scene_dropdown, response])
|
366 |
sample_dropdown.change(update2, [sample_dropdown, scene_dropdown, prompt_type_dropdown], [prompt_sample_dropdown, response])
|
367 |
scene_dropdown.change(update2, [sample_dropdown, scene_dropdown, prompt_type_dropdown], [prompt_sample_dropdown, response])
|
368 |
prompt_type_dropdown.change(update2, [sample_dropdown, scene_dropdown, prompt_type_dropdown], [prompt_sample_dropdown, response])
|
369 |
-
|
370 |
-
# sample_dropdown.change(update1, sample_dropdown, [scene_dropdown, response, vx])
|
371 |
-
# sample_dropdown.change(update2, [sample_dropdown, scene_dropdown, prompt_type_dropdown], [prompt_sample_dropdown, response, vx])
|
372 |
-
# scene_dropdown.change(update2, [sample_dropdown, scene_dropdown, prompt_type_dropdown], [prompt_sample_dropdown, response, vx])
|
373 |
-
# prompt_type_dropdown.change(update2, [sample_dropdown, scene_dropdown, prompt_type_dropdown], [prompt_sample_dropdown, response, vx])
|
374 |
-
# prompt_sample_dropdown.change(update3, [sample_dropdown, scene_dropdown, prompt_type_dropdown, prompt_sample_dropdown], [response, vx])
|
375 |
-
|
376 |
-
# Logic to handle interactions
|
377 |
show_button.click(load_3d_scene, inputs=[sample_dropdown, scene_dropdown], outputs=plot1)
|
378 |
-
show_prompt_button.click(show_prompt_in_3d, inputs=[sample_dropdown, scene_dropdown, prompt_type_dropdown, prompt_sample_dropdown], outputs=plot1)
|
379 |
-
# start_segment_button.click(start_segmentation, inputs=[sample_dropdown, scene_dropdown, prompt_type_dropdown, prompt_sample_dropdown, vx], outputs=[plot1, response])
|
380 |
start_segment_button.click(start_segmentation, inputs=[sample_dropdown, scene_dropdown, prompt_type_dropdown, prompt_sample_dropdown], outputs=[plot1, response])
|
381 |
|
382 |
app.queue(status_update_rate="auto")
|
383 |
app.launch(share=True, favicon_path="./logo.png")
|
384 |
|
385 |
-
|
386 |
if __name__ == "__main__":
|
387 |
-
main()
|
388 |
-
|
389 |
-
|
390 |
-
|
391 |
-
|
392 |
-
|
393 |
-
|
|
|
5 |
from sam2point import dataset
|
6 |
import sam2point.configs as configs
|
7 |
from demo_utils import run_demo, create_box
|
8 |
+
import spaces
|
9 |
+
|
10 |
samples = {
|
11 |
"3D Indoor Scene - S3DIS": ["Conference Room", "Restroom", "Lobby", "Office1", "Office2"],
|
|
|
12 |
"3D Indoor Scene - ScanNet": ["Scene1", "Scene2", "Scene3", "Scene4", "Scene5", "Scene6"],
|
13 |
"3D Outdoor Driving Scene - KITTI": ["Scene1", "Scene2", "Scene3", "Scene4", "Scene5", "Scene6"],
|
14 |
"3D Outdoor Street Scene - Semantic3D": ["Scene1", "Scene2", "Scene3", "Scene4", "Scene5", "Scene6", "Scene7"],
|
15 |
"3D Object - Objaverse": ["Plant", "Lego", "Lock", "Eleplant", "Knife Rest", "Skateboard", "Popcorn Machine", "Stove", "Bus Shelter", "Thor Hammer", "Horse"],
|
|
|
16 |
}
|
17 |
|
|
|
18 |
PATH = {
|
19 |
"S3DIS": ['Area_1_conferenceRoom_1.txt', 'Area_2_WC_1.txt', 'Area_4_lobby_2.txt', 'Area_5_office_3.txt', 'Area_6_office_9.txt'],
|
|
|
20 |
"ScanNet": ['scene0005_01.pth', 'scene0010_01.pth', 'scene0016_02.pth', 'scene0019_01.pth', 'scene0000_00.pth', 'scene0002_00.pth'],
|
21 |
"Objaverse": ["plant.npy", "human.npy", "lock.npy", "elephant.npy", "knife_rest.npy", "skateboard.npy", "popcorn_machine.npy", "stove.npy", "bus_shelter.npy", "thor_hammer.npy", "horse.npy"],
|
|
|
22 |
"KITTI": ["scene1.npy", "scene2.npy", "scene3.npy", "scene4.npy", "scene5.npy", "scene6.npy"],
|
23 |
"Semantic3D": ["scene1.npy", "scene2.npy", "patch19.npy", "patch0.npy", "patch1.npy", "patch50.npy", "patch62.npy"]
|
24 |
}
|
25 |
|
|
|
26 |
prompt_types = ["Point", "Box", "Mask"]
|
27 |
|
28 |
+
@spaces.GPU()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
29 |
def load_3d_scene(name, sample_idx=-1, type_=None, prompt=None, final=False, new_color=None):
|
30 |
DATASET = name.split('-')[1].replace(" ", "")
|
31 |
path = 'data/' + DATASET + '/' + PATH[DATASET][sample_idx]
|
32 |
asp, SIZE = 1., 1
|
33 |
+
|
34 |
print(path)
|
35 |
if DATASET == 'S3DIS':
|
36 |
point, color = dataset.load_S3DIS_sample(path, sample=True)
|
|
|
49 |
elif DATASET == 'Semantic3D':
|
50 |
point, color = dataset.load_Semantic3D_sample(path, sample_idx, sample=True)
|
51 |
alpha = 0.2
|
52 |
+
print("Loading Dataset:", DATASET, "Point Cloud Size:", point.shape, "Path:", path)
|
53 |
|
54 |
+
##### Initial Show #####
|
|
|
55 |
if not type_:
|
56 |
+
if point.shape[0] > 100000: # sample points for speeding up
|
57 |
indices = np.random.choice(point.shape[0], 100000, replace=False)
|
58 |
point = point[indices]
|
59 |
color = color[indices]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
60 |
fig = go.Figure(
|
61 |
data=[
|
62 |
go.Scatter3d(
|
|
|
77 |
)
|
78 |
)
|
79 |
return fig
|
80 |
+
##### Final Results #####
|
81 |
if final:
|
82 |
color = new_color
|
83 |
green = np.array([[0.1, 0.1, 0.1]])
|
|
|
92 |
indices = np.random.choice(point.shape[0], 100000, replace=False)
|
93 |
point = point[indices]
|
94 |
color = color[indices]
|
|
|
|
|
|
|
|
|
95 |
scatter = go.Scatter3d(
|
96 |
x=point[:,0], y=point[:,1], z=point[:,2],
|
97 |
mode='markers',
|
|
|
100 |
)
|
101 |
if final: scatter = [scatter, add_green] + create_box(prompt)
|
102 |
else: scatter = [scatter] + create_box(prompt)
|
|
|
103 |
elif type_ == "point":
|
104 |
prompt = np.array([prompt])
|
105 |
new = go.Scatter3d(
|
106 |
x=prompt[:,0], y=prompt[:,1], z=prompt[:,2],
|
107 |
mode='markers',
|
108 |
+
marker=dict(size=5, color='red', opacity=1),
|
|
|
|
|
109 |
name="Point Prompt"
|
110 |
)
|
|
|
111 |
if point.shape[0] > 100000:
|
112 |
indices = np.random.choice(point.shape[0], 100000, replace=False)
|
113 |
point = point[indices]
|
114 |
color = color[indices]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
115 |
scatter = go.Scatter3d(
|
116 |
x=point[:,0], y=point[:,1], z=point[:,2],
|
117 |
mode='markers',
|
|
|
145 |
indices = np.random.choice(point.shape[0], 100000, replace=False)
|
146 |
point = point[indices]
|
147 |
color = color[indices]
|
|
|
|
|
|
|
|
|
|
|
|
|
148 |
scatter = go.Scatter3d(
|
149 |
x=point[:,0], y=point[:,1], z=point[:,2],
|
150 |
mode='markers',
|
|
|
152 |
name="3D Object/Scene"
|
153 |
)
|
154 |
scatter = [scatter, add_green]
|
|
|
155 |
else:
|
156 |
print("Wrong Prompt Type")
|
157 |
exit(1)
|
158 |
|
|
|
159 |
fig = go.Figure(
|
160 |
data=scatter,
|
161 |
layout=dict(
|
|
|
170 |
)
|
171 |
return fig
|
172 |
|
173 |
+
@spaces.GPU()
|
|
|
|
|
|
|
174 |
def show_prompt_in_3d(name, sample_idx, prompt_type, prompt_idx):
|
175 |
+
if name == None or sample_idx == None or prompt_type == None or prompt_idx == None:
|
176 |
+
return gr.Plot(), gr.Textbox(label="Response", value="Please ensure all options are selected.", visible=True)
|
177 |
+
|
178 |
DATASET = name.split('-')[1].replace(" ", "")
|
179 |
TYPE = prompt_type.lower()
|
180 |
theta = 0. if DATASET in "S3DIS ScanNet" else 0.5
|
181 |
mode = "bilinear" if DATASET in "S3DIS ScanNet" else 'nearest'
|
182 |
+
|
|
|
183 |
prompt = run_demo(DATASET, TYPE, sample_idx, prompt_idx, 0.02, theta, mode, ret_prompt=True)
|
184 |
fig = load_3d_scene(name, sample_idx, TYPE, prompt)
|
185 |
+
return fig, gr.Textbox(label="Response", value="Prompt has been shown in 3D Object/Scene!", visible=True)
|
|
|
186 |
|
187 |
+
@spaces.GPU()
|
|
|
|
|
188 |
def start_segmentation(name=None, sample_idx=None, prompt_type=None, prompt_idx=None, vx=0.02):
|
189 |
if name == None or sample_idx == None or prompt_type == None or prompt_idx == None:
|
190 |
return gr.Plot(), gr.Textbox(label="Response", value="Please ensure all options are selected.", visible=True)
|
|
|
194 |
theta = 0. if DATASET in "S3DIS ScanNet" else 0.5
|
195 |
mode = "bilinear" if DATASET in "S3DIS ScanNet" else 'nearest'
|
196 |
|
|
|
197 |
new_color, prompt = run_demo(DATASET, TYPE, sample_idx, prompt_idx, vx, theta, mode, ret_prompt=False)
|
198 |
fig = load_3d_scene(name, sample_idx, TYPE, prompt, final=True, new_color=new_color)
|
199 |
return fig, gr.Textbox(label="Response", value="Segmentation completed successfully!", visible=True)
|
200 |
|
201 |
+
@spaces.GPU()
|
|
|
|
|
202 |
def update1(datasets):
|
203 |
if 'Objaverse' in datasets:
|
204 |
+
return gr.Radio(label="Select 3D Object", choices=samples[datasets]), gr.Textbox(label="Response", value="", visible=True)
|
205 |
+
return gr.Radio(label="Select 3D Scene", choices=samples[datasets]), gr.Textbox(label="Response", value="", visible=True)
|
|
|
206 |
|
207 |
+
@spaces.GPU()
|
208 |
def update2(name, sample_idx, prompt_type):
|
209 |
if name == None or sample_idx == None or prompt_type == None:
|
210 |
+
return gr.Radio(label="Select Prompt Example", choices=[]), gr.Textbox(label="Response", value="", visible=True)
|
211 |
DATASET = name.split('-')[1].replace(" ", "")
|
212 |
TYPE = prompt_type.lower() + '_prompts'
|
213 |
+
|
214 |
if DATASET == 'S3DIS':
|
215 |
info = configs.S3DIS_samples[sample_idx][TYPE]
|
216 |
elif DATASET == 'ScanNet':
|
|
|
223 |
info = configs.Semantic3D_samples[sample_idx][TYPE]
|
224 |
|
225 |
cur = ['Example ' + str(i) for i in range(1, len(info) + 1)]
|
226 |
+
return gr.Radio(label="Select Prompt Example", choices=cur), gr.Textbox(label="Response", value="", visible=True)
|
227 |
+
|
228 |
+
@spaces.GPU()
|
229 |
def update3(name, sample_idx, prompt_type, prompt_idx):
|
230 |
if name == None or sample_idx == None or prompt_type == None:
|
231 |
return gr.Textbox(label="Response", value="", visible=True), gr.Slider(minimum=0.01, maximum=0.15, step=0.001, label="Voxel Size", value=0.02)
|
232 |
DATASET = name.split('-')[1].replace(" ", "")
|
233 |
TYPE = configs.VOXEL[prompt_type.lower()]
|
234 |
+
|
235 |
if DATASET in "S3DIS ScanNet":
|
236 |
vx_ = 0.02
|
237 |
elif DATASET == 'Objaverse':
|
|
|
243 |
|
244 |
return gr.Textbox(label="Response", value="", visible=True), gr.Slider(minimum=0.01, maximum=0.15, step=0.001, label="Voxel Size", value=vx_)
|
245 |
|
246 |
+
@spaces.GPU()
|
247 |
def main():
|
248 |
+
title = """<h1 style="text-align: center;">
|
|
|
|
|
|
|
|
|
|
|
249 |
<div style="width: 1.2em; height: 1.2em; display: inline-block;"><img src="https://github.com/ZiyuGuo99/ZiyuGuo99.github.io/blob/main/assets/img/logo.png?raw=true" style='width: 100%; height: 100%; object-fit: contain;' /></div>
|
250 |
<span style="font-variant: small-caps; font-weight: bold;">Sam2Point</span>
|
251 |
</h1>
|
|
|
286 |
prompt_type_dropdown = gr.Radio(label="Select Prompt Type", choices=prompt_types)
|
287 |
prompt_sample_dropdown = gr.Radio(label="Select Prompt Example", choices=[], type="index")
|
288 |
show_prompt_button = gr.Button("Show Prompt in 3D Scene/Object")
|
|
|
289 |
with gr.Column():
|
|
|
290 |
start_segment_button = gr.Button("Start Segmentation")
|
291 |
plot1 = gr.Plot()
|
292 |
|
|
|
|
|
|
|
293 |
response = gr.Textbox(label="Response")
|
294 |
|
295 |
sample_dropdown.change(update1, sample_dropdown, [scene_dropdown, response])
|
296 |
sample_dropdown.change(update2, [sample_dropdown, scene_dropdown, prompt_type_dropdown], [prompt_sample_dropdown, response])
|
297 |
scene_dropdown.change(update2, [sample_dropdown, scene_dropdown, prompt_type_dropdown], [prompt_sample_dropdown, response])
|
298 |
prompt_type_dropdown.change(update2, [sample_dropdown, scene_dropdown, prompt_type_dropdown], [prompt_sample_dropdown, response])
|
299 |
+
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
300 |
show_button.click(load_3d_scene, inputs=[sample_dropdown, scene_dropdown], outputs=plot1)
|
301 |
+
show_prompt_button.click(show_prompt_in_3d, inputs=[sample_dropdown, scene_dropdown, prompt_type_dropdown, prompt_sample_dropdown], outputs=[plot1, response])
|
|
|
302 |
start_segment_button.click(start_segmentation, inputs=[sample_dropdown, scene_dropdown, prompt_type_dropdown, prompt_sample_dropdown], outputs=[plot1, response])
|
303 |
|
304 |
app.queue(status_update_rate="auto")
|
305 |
app.launch(share=True, favicon_path="./logo.png")
|
306 |
|
|
|
307 |
if __name__ == "__main__":
|
308 |
+
main()
|
|
|
|
|
|
|
|
|
|
|
|