Spaces:
Sleeping
Sleeping
fix: Update app.py
Browse files
app.py
CHANGED
@@ -3,6 +3,7 @@ import time
|
|
3 |
import torch
|
4 |
import numpy as np
|
5 |
import gradio as gr
|
|
|
6 |
import tempfile
|
7 |
import subprocess
|
8 |
from dust3r.losses import L21
|
@@ -11,6 +12,7 @@ from spann3r.datasets import Demo
|
|
11 |
from torch.utils.data import DataLoader
|
12 |
import trimesh
|
13 |
from scipy.spatial.transform import Rotation
|
|
|
14 |
|
15 |
# Default values
|
16 |
DEFAULT_CKPT_PATH = './checkpoints/spann3r.pth'
|
@@ -45,10 +47,22 @@ def cat_meshes(meshes):
|
|
45 |
faces = np.concatenate(faces)
|
46 |
return dict(vertices=vertices, face_colors=colors, faces=faces)
|
47 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
48 |
def load_model(ckpt_path, device):
|
49 |
model = Spann3R(dus3r_name=DEFAULT_DUST3R_PATH,
|
50 |
use_feat=False).to(device)
|
51 |
-
|
|
|
52 |
model.eval()
|
53 |
return model
|
54 |
|
@@ -91,14 +105,14 @@ def pts3d_to_trimesh(img, pts3d, valid=None):
|
|
91 |
assert len(faces) == len(face_colors)
|
92 |
return dict(vertices=vertices, face_colors=face_colors, faces=faces)
|
93 |
|
|
|
|
|
|
|
94 |
@torch.no_grad()
|
95 |
-
def reconstruct(video_path, conf_thresh, kf_every,
|
96 |
# Extract frames from video
|
97 |
demo_path = extract_frames(video_path)
|
98 |
|
99 |
-
# Load model
|
100 |
-
model = load_model(DEFAULT_CKPT_PATH, DEFAULT_DEVICE)
|
101 |
-
|
102 |
# Load dataset
|
103 |
dataset = Demo(ROOT=demo_path, resolution=224, full_video=True, kf_every=kf_every)
|
104 |
dataloader = DataLoader(dataset, batch_size=1, shuffle=False, num_workers=0)
|
@@ -168,16 +182,15 @@ iface = gr.Interface(
|
|
168 |
inputs=[
|
169 |
gr.Video(label="Input Video"),
|
170 |
gr.Slider(0, 1, value=1e-3, label="Confidence Threshold"),
|
171 |
-
gr.Slider(1, 30, step=1, value=
|
172 |
-
gr.Slider(0.001, 0.01, value=0.005, step=0.001, label="Voxel Size for Downsampling"),
|
173 |
gr.Checkbox(label="As Pointcloud", value=False)
|
174 |
],
|
175 |
outputs=[
|
176 |
gr.Model3D(label="3D Model (GLB)", display_mode="solid"),
|
177 |
gr.Textbox(label="Status")
|
178 |
],
|
179 |
-
title="3D Reconstruction
|
180 |
)
|
181 |
|
182 |
if __name__ == "__main__":
|
183 |
-
iface.launch(
|
|
|
3 |
import torch
|
4 |
import numpy as np
|
5 |
import gradio as gr
|
6 |
+
import urllib.parse
|
7 |
import tempfile
|
8 |
import subprocess
|
9 |
from dust3r.losses import L21
|
|
|
12 |
from torch.utils.data import DataLoader
|
13 |
import trimesh
|
14 |
from scipy.spatial.transform import Rotation
|
15 |
+
import spaces
|
16 |
|
17 |
# Default values
|
18 |
DEFAULT_CKPT_PATH = './checkpoints/spann3r.pth'
|
|
|
47 |
faces = np.concatenate(faces)
|
48 |
return dict(vertices=vertices, face_colors=colors, faces=faces)
|
49 |
|
50 |
+
def load_ckpt(model_path_or_url, verbose=True):
|
51 |
+
if verbose:
|
52 |
+
print('... loading model from', model_path_or_url)
|
53 |
+
is_url = urllib.parse.urlparse(model_path_or_url).scheme in ('http', 'https')
|
54 |
+
|
55 |
+
if is_url:
|
56 |
+
ckpt = torch.hub.load_state_dict_from_url(model_path_or_url, map_location='cpu', progress=verbose)
|
57 |
+
else:
|
58 |
+
ckpt = torch.load(model_path_or_url, map_location='cpu')
|
59 |
+
return ckpt
|
60 |
+
|
61 |
def load_model(ckpt_path, device):
|
62 |
model = Spann3R(dus3r_name=DEFAULT_DUST3R_PATH,
|
63 |
use_feat=False).to(device)
|
64 |
+
|
65 |
+
model.load_state_dict(load_ckpt(ckpt_path)['model'])
|
66 |
model.eval()
|
67 |
return model
|
68 |
|
|
|
105 |
assert len(faces) == len(face_colors)
|
106 |
return dict(vertices=vertices, face_colors=face_colors, faces=faces)
|
107 |
|
108 |
+
model = load_model(DEFAULT_CKPT_PATH, DEFAULT_DEVICE)
|
109 |
+
|
110 |
+
@spaces.GPU
|
111 |
@torch.no_grad()
|
112 |
+
def reconstruct(video_path, conf_thresh, kf_every, as_pointcloud=False):
|
113 |
# Extract frames from video
|
114 |
demo_path = extract_frames(video_path)
|
115 |
|
|
|
|
|
|
|
116 |
# Load dataset
|
117 |
dataset = Demo(ROOT=demo_path, resolution=224, full_video=True, kf_every=kf_every)
|
118 |
dataloader = DataLoader(dataset, batch_size=1, shuffle=False, num_workers=0)
|
|
|
182 |
inputs=[
|
183 |
gr.Video(label="Input Video"),
|
184 |
gr.Slider(0, 1, value=1e-3, label="Confidence Threshold"),
|
185 |
+
gr.Slider(1, 30, step=1, value=5, label="Keyframe Interval"),
|
|
|
186 |
gr.Checkbox(label="As Pointcloud", value=False)
|
187 |
],
|
188 |
outputs=[
|
189 |
gr.Model3D(label="3D Model (GLB)", display_mode="solid"),
|
190 |
gr.Textbox(label="Status")
|
191 |
],
|
192 |
+
title="3D Reconstruction with Spatial Memory",
|
193 |
)
|
194 |
|
195 |
if __name__ == "__main__":
|
196 |
+
iface.launch()
|