Spaces:
Runtime error
Runtime error
liuyuan-pal
commited on
Commit
•
df916e6
1
Parent(s):
a14768e
update
Browse files- .gitattributes +1 -0
- app.py +26 -4
- detection_test.py +30 -9
- hf_demo/examples/basket.png +0 -0
- hf_demo/examples/cat.png +3 -0
- hf_demo/examples/crab.png +3 -0
- hf_demo/examples/elephant.png +3 -0
- hf_demo/examples/flower.png +3 -0
- hf_demo/examples/forest.png +3 -0
- hf_demo/examples/monkey.png +0 -0
- hf_demo/examples/teapot.png +3 -0
.gitattributes
CHANGED
@@ -34,3 +34,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
36 |
ckpt/* filter=lfs diff=lfs merge=lfs -text
|
|
|
|
34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
36 |
ckpt/* filter=lfs diff=lfs merge=lfs -text
|
37 |
+
hf_demo/examples/* filter=lfs diff=lfs merge=lfs -text
|
app.py
CHANGED
@@ -7,7 +7,6 @@ import torch
|
|
7 |
import os
|
8 |
import fire
|
9 |
from omegaconf import OmegaConf
|
10 |
-
from rembg import remove
|
11 |
|
12 |
from ldm.util import add_margin, instantiate_from_config
|
13 |
from sam_utils import sam_init, sam_out_nosave
|
@@ -39,6 +38,28 @@ _USER_GUIDE3 = "Generated multiview images are shown below!"
|
|
39 |
|
40 |
deployed = True
|
41 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
42 |
def resize_inputs(image_input, crop_size):
|
43 |
alpha_np = np.asarray(image_input)[:, :, 3]
|
44 |
coords = np.stack(np.nonzero(alpha_np), 1)[:, (1, 0)]
|
@@ -95,9 +116,9 @@ def white_background(img):
|
|
95 |
rgb = (rgb*255).astype(np.uint8)
|
96 |
return Image.fromarray(rgb)
|
97 |
|
98 |
-
def sam_predict(predictor, raw_im):
|
99 |
raw_im.thumbnail([512, 512], Image.Resampling.LANCZOS)
|
100 |
-
image_nobg =
|
101 |
arr = np.asarray(image_nobg)[:, :, -1]
|
102 |
x_nonzero = np.nonzero(arr.sum(axis=0))
|
103 |
y_nonzero = np.nonzero(arr.sum(axis=1))
|
@@ -140,6 +161,7 @@ def run_demo():
|
|
140 |
|
141 |
# init sam model
|
142 |
mask_predictor = sam_init()
|
|
|
143 |
|
144 |
# with open('instructions_12345.md', 'r') as f:
|
145 |
# article = f.read()
|
@@ -192,7 +214,7 @@ def run_demo():
|
|
192 |
output_block = gr.Image(type='pil', image_mode='RGB', label="Outputs of SyncDreamer", height=256, interactive=False)
|
193 |
|
194 |
update_guide = lambda GUIDE_TEXT: gr.update(value=GUIDE_TEXT)
|
195 |
-
image_block.change(fn=partial(sam_predict, mask_predictor), inputs=[image_block], outputs=[sam_block], queue=False)\
|
196 |
.success(fn=partial(update_guide, _USER_GUIDE1), outputs=[guide_text], queue=False)
|
197 |
|
198 |
crop_size_slider.change(fn=resize_inputs, inputs=[sam_block, crop_size_slider], outputs=[input_block], queue=False)\
|
|
|
7 |
import os
|
8 |
import fire
|
9 |
from omegaconf import OmegaConf
|
|
|
10 |
|
11 |
from ldm.util import add_margin, instantiate_from_config
|
12 |
from sam_utils import sam_init, sam_out_nosave
|
|
|
38 |
|
39 |
deployed = True
|
40 |
|
41 |
+
class BackgroundRemoval:
|
42 |
+
def __init__(self, device='cuda'):
|
43 |
+
from carvekit.api.high import HiInterface
|
44 |
+
self.interface = HiInterface(
|
45 |
+
object_type="object", # Can be "object" or "hairs-like".
|
46 |
+
batch_size_seg=5,
|
47 |
+
batch_size_matting=1,
|
48 |
+
device=device,
|
49 |
+
seg_mask_size=640, # Use 640 for Tracer B7 and 320 for U2Net
|
50 |
+
matting_mask_size=2048,
|
51 |
+
trimap_prob_threshold=231,
|
52 |
+
trimap_dilation=30,
|
53 |
+
trimap_erosion_iters=5,
|
54 |
+
fp16=True,
|
55 |
+
)
|
56 |
+
|
57 |
+
@torch.no_grad()
|
58 |
+
def __call__(self, image):
|
59 |
+
# image: [H, W, 3] array in [0, 255].
|
60 |
+
image = self.interface([image])[0]
|
61 |
+
return image
|
62 |
+
|
63 |
def resize_inputs(image_input, crop_size):
|
64 |
alpha_np = np.asarray(image_input)[:, :, 3]
|
65 |
coords = np.stack(np.nonzero(alpha_np), 1)[:, (1, 0)]
|
|
|
116 |
rgb = (rgb*255).astype(np.uint8)
|
117 |
return Image.fromarray(rgb)
|
118 |
|
119 |
+
def sam_predict(predictor, removal, raw_im):
|
120 |
raw_im.thumbnail([512, 512], Image.Resampling.LANCZOS)
|
121 |
+
image_nobg = removal(raw_im.convert('RGB'))
|
122 |
arr = np.asarray(image_nobg)[:, :, -1]
|
123 |
x_nonzero = np.nonzero(arr.sum(axis=0))
|
124 |
y_nonzero = np.nonzero(arr.sum(axis=1))
|
|
|
161 |
|
162 |
# init sam model
|
163 |
mask_predictor = sam_init()
|
164 |
+
removal = BackgroundRemoval()
|
165 |
|
166 |
# with open('instructions_12345.md', 'r') as f:
|
167 |
# article = f.read()
|
|
|
214 |
output_block = gr.Image(type='pil', image_mode='RGB', label="Outputs of SyncDreamer", height=256, interactive=False)
|
215 |
|
216 |
update_guide = lambda GUIDE_TEXT: gr.update(value=GUIDE_TEXT)
|
217 |
+
image_block.change(fn=partial(sam_predict, mask_predictor, removal), inputs=[image_block], outputs=[sam_block], queue=False)\
|
218 |
.success(fn=partial(update_guide, _USER_GUIDE1), outputs=[guide_text], queue=False)
|
219 |
|
220 |
crop_size_slider.change(fn=resize_inputs, inputs=[sam_block, crop_size_slider], outputs=[input_block], queue=False)\
|
detection_test.py
CHANGED
@@ -1,18 +1,39 @@
|
|
|
|
1 |
import numpy as np
|
2 |
from PIL import Image
|
3 |
from skimage.io import imsave
|
4 |
-
|
5 |
-
from app import white_background
|
6 |
-
from ldm.util import add_margin
|
7 |
from sam_utils import sam_out_nosave, sam_init
|
8 |
-
from rembg import remove
|
9 |
|
10 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
11 |
predictor = sam_init()
|
12 |
|
13 |
raw_im.thumbnail([512, 512], Image.Resampling.LANCZOS)
|
14 |
width, height = raw_im.size
|
15 |
-
image_nobg =
|
16 |
arr = np.asarray(image_nobg)[:, :, -1]
|
17 |
x_nonzero = np.nonzero(arr.sum(axis=0))
|
18 |
y_nonzero = np.nonzero(arr.sum(axis=1))
|
@@ -20,16 +41,16 @@ x_min = int(x_nonzero[0].min())
|
|
20 |
y_min = int(y_nonzero[0].min())
|
21 |
x_max = int(x_nonzero[0].max())
|
22 |
y_max = int(y_nonzero[0].max())
|
23 |
-
|
24 |
|
25 |
image_nobg.thumbnail([512, 512], Image.Resampling.LANCZOS)
|
26 |
image_sam = sam_out_nosave(predictor, image_nobg.convert("RGB"), (x_min, y_min, x_max, y_max))
|
27 |
|
28 |
-
|
29 |
image_sam = np.asarray(image_sam, np.float32) / 255
|
30 |
out_mask = image_sam[:, :, 3:]
|
31 |
out_rgb = image_sam[:, :, :3] * out_mask + 1 - out_mask
|
32 |
out_img = (np.concatenate([out_rgb, out_mask], 2) * 255).astype(np.uint8)
|
33 |
|
34 |
image_sam = Image.fromarray(out_img, mode='RGBA')
|
35 |
-
|
|
|
1 |
+
import torch
|
2 |
import numpy as np
|
3 |
from PIL import Image
|
4 |
from skimage.io import imsave
|
|
|
|
|
|
|
5 |
from sam_utils import sam_out_nosave, sam_init
|
|
|
6 |
|
7 |
+
class BackgroundRemoval:
|
8 |
+
def __init__(self, device='cuda'):
|
9 |
+
from carvekit.api.high import HiInterface
|
10 |
+
self.interface = HiInterface(
|
11 |
+
object_type="object", # Can be "object" or "hairs-like".
|
12 |
+
batch_size_seg=5,
|
13 |
+
batch_size_matting=1,
|
14 |
+
device=device,
|
15 |
+
seg_mask_size=640, # Use 640 for Tracer B7 and 320 for U2Net
|
16 |
+
matting_mask_size=2048,
|
17 |
+
trimap_prob_threshold=231,
|
18 |
+
trimap_dilation=30,
|
19 |
+
trimap_erosion_iters=5,
|
20 |
+
fp16=True,
|
21 |
+
)
|
22 |
+
|
23 |
+
@torch.no_grad()
|
24 |
+
def __call__(self, image):
|
25 |
+
# image: [H, W, 3] array in [0, 255].
|
26 |
+
# image = Image.fromarray(image)
|
27 |
+
image = self.interface([image])[0]
|
28 |
+
# image = np.array(image)
|
29 |
+
return image
|
30 |
+
|
31 |
+
raw_im = Image.open('hf_demo/examples/flower.png')
|
32 |
predictor = sam_init()
|
33 |
|
34 |
raw_im.thumbnail([512, 512], Image.Resampling.LANCZOS)
|
35 |
width, height = raw_im.size
|
36 |
+
image_nobg = BackgroundRemoval()(raw_im.convert('RGB'))
|
37 |
arr = np.asarray(image_nobg)[:, :, -1]
|
38 |
x_nonzero = np.nonzero(arr.sum(axis=0))
|
39 |
y_nonzero = np.nonzero(arr.sum(axis=1))
|
|
|
41 |
y_min = int(y_nonzero[0].min())
|
42 |
x_max = int(x_nonzero[0].max())
|
43 |
y_max = int(y_nonzero[0].max())
|
44 |
+
image_nobg.save('./nobg.png')
|
45 |
|
46 |
image_nobg.thumbnail([512, 512], Image.Resampling.LANCZOS)
|
47 |
image_sam = sam_out_nosave(predictor, image_nobg.convert("RGB"), (x_min, y_min, x_max, y_max))
|
48 |
|
49 |
+
imsave('./mask.png', np.asarray(image_sam)[:,:,3])
|
50 |
image_sam = np.asarray(image_sam, np.float32) / 255
|
51 |
out_mask = image_sam[:, :, 3:]
|
52 |
out_rgb = image_sam[:, :, :3] * out_mask + 1 - out_mask
|
53 |
out_img = (np.concatenate([out_rgb, out_mask], 2) * 255).astype(np.uint8)
|
54 |
|
55 |
image_sam = Image.fromarray(out_img, mode='RGBA')
|
56 |
+
image_sam.save('./output.png')
|
hf_demo/examples/basket.png
CHANGED
Git LFS Details
|
hf_demo/examples/cat.png
ADDED
Git LFS Details
|
hf_demo/examples/crab.png
ADDED
Git LFS Details
|
hf_demo/examples/elephant.png
ADDED
Git LFS Details
|
hf_demo/examples/flower.png
ADDED
Git LFS Details
|
hf_demo/examples/forest.png
ADDED
Git LFS Details
|
hf_demo/examples/monkey.png
CHANGED
Git LFS Details
|
hf_demo/examples/teapot.png
ADDED
Git LFS Details
|