Spaces:
Runtime error
Runtime error
Plat
commited on
Commit
·
3e0e479
1
Parent(s):
5bffca4
fix
Browse files- .gitignore +2 -0
- .python-version +1 -0
- app.py +30 -22
- pyproject.toml +20 -0
- requirements.txt +1 -0
- uv.lock +0 -0
.gitignore
ADDED
@@ -0,0 +1,2 @@
|
|
|
|
|
|
|
1 |
+
.venv
|
2 |
+
.mypy_cache
|
.python-version
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
3.11
|
app.py
CHANGED
@@ -1,13 +1,14 @@
|
|
1 |
import gradio as gr
|
2 |
import numpy as np
|
3 |
-
from
|
|
|
4 |
|
5 |
import torch
|
6 |
-
import torchvision.
|
7 |
|
8 |
|
9 |
MODEL_MAP = {
|
10 |
-
"
|
11 |
"mean": [0, 0, 0],
|
12 |
"std": [1.0, 1.0, 1.0],
|
13 |
"image_size": 384,
|
@@ -15,9 +16,11 @@ MODEL_MAP = {
|
|
15 |
}
|
16 |
}
|
17 |
|
|
|
18 |
def config_to_processor(config: dict):
|
19 |
return T.Compose(
|
20 |
[
|
|
|
21 |
T.Resize(
|
22 |
size=None,
|
23 |
max_size=config["image_size"],
|
@@ -25,55 +28,60 @@ def config_to_processor(config: dict):
|
|
25 |
),
|
26 |
T.Pad(
|
27 |
padding=config["image_size"] // 2,
|
28 |
-
fill=config["background]
|
29 |
),
|
30 |
T.CenterCrop(
|
31 |
size=(config["image_size"], config["image_size"]),
|
32 |
),
|
33 |
-
T.
|
34 |
-
T.ToDtype(dtype=torch.float32, scale=True), # 0~255 -> 0~1
|
35 |
T.Normalize(mean=config["mean"], std=config["std"]),
|
36 |
]
|
37 |
)
|
38 |
|
|
|
39 |
def load_model(name: str):
|
40 |
-
return
|
|
|
41 |
|
42 |
MODELS = {
|
43 |
name: {
|
44 |
"model": load_model(name),
|
45 |
"processor": config_to_processor(config),
|
46 |
}
|
47 |
-
for name, config in
|
48 |
}
|
49 |
|
50 |
|
51 |
@torch.inference_mode()
|
52 |
-
def calculate_similarity(
|
53 |
model = MODELS[model_name]["model"]
|
54 |
processor = MODELS[model_name]["processor"]
|
55 |
-
|
56 |
-
pixel_values = torch.
|
57 |
-
|
58 |
-
|
59 |
-
embeddings = model(pixel_values)
|
60 |
embeddings /= embeddings.norm(p=2, dim=-1, keepdim=True)
|
61 |
|
62 |
-
similarity = (embeddings[0]
|
|
|
63 |
return similarity
|
64 |
-
|
65 |
|
66 |
with gr.Blocks() as demo:
|
67 |
with gr.Row():
|
68 |
with gr.Column():
|
69 |
-
image_1 = gr.Image("Image 1", type="pil")
|
70 |
-
image_2 = gr.Image("Image 2", type="pil")
|
71 |
|
72 |
-
model_name = gr.
|
|
|
|
|
|
|
|
|
73 |
submit_btn = gr.Button("Submit", variant="primary")
|
74 |
-
|
75 |
with gr.Column():
|
76 |
-
similarity = gr.
|
77 |
|
78 |
gr.on(
|
79 |
triggers=[submit_btn.click],
|
@@ -83,7 +91,7 @@ with gr.Blocks() as demo:
|
|
83 |
image_1,
|
84 |
image_2,
|
85 |
],
|
86 |
-
outputs=[
|
87 |
)
|
88 |
|
89 |
if __name__ == "__main__":
|
|
|
1 |
import gradio as gr
|
2 |
import numpy as np
|
3 |
+
from PIL import Image
|
4 |
+
from transformers import TimmWrapperModel
|
5 |
|
6 |
import torch
|
7 |
+
import torchvision.transforms.v2 as T
|
8 |
|
9 |
|
10 |
MODEL_MAP = {
|
11 |
+
"p1atdev/style_250412.vit_base_patch16_siglip_384.v2_webli": {
|
12 |
"mean": [0, 0, 0],
|
13 |
"std": [1.0, 1.0, 1.0],
|
14 |
"image_size": 384,
|
|
|
16 |
}
|
17 |
}
|
18 |
|
19 |
+
|
20 |
def config_to_processor(config: dict):
|
21 |
return T.Compose(
|
22 |
[
|
23 |
+
T.PILToTensor(),
|
24 |
T.Resize(
|
25 |
size=None,
|
26 |
max_size=config["image_size"],
|
|
|
28 |
),
|
29 |
T.Pad(
|
30 |
padding=config["image_size"] // 2,
|
31 |
+
fill=config["background"],
|
32 |
),
|
33 |
T.CenterCrop(
|
34 |
size=(config["image_size"], config["image_size"]),
|
35 |
),
|
36 |
+
T.ToDtype(dtype=torch.float32, scale=True), # 0~255 -> 0~1
|
|
|
37 |
T.Normalize(mean=config["mean"], std=config["std"]),
|
38 |
]
|
39 |
)
|
40 |
|
41 |
+
|
42 |
def load_model(name: str):
|
43 |
+
return TimmWrapperModel.from_pretrained(name).eval().requires_grad_(False)
|
44 |
+
|
45 |
|
46 |
MODELS = {
|
47 |
name: {
|
48 |
"model": load_model(name),
|
49 |
"processor": config_to_processor(config),
|
50 |
}
|
51 |
+
for name, config in MODEL_MAP.items()
|
52 |
}
|
53 |
|
54 |
|
55 |
@torch.inference_mode()
|
56 |
+
def calculate_similarity(model_name: str, image_1: Image.Image, image_2: Image.Image):
|
57 |
model = MODELS[model_name]["model"]
|
58 |
processor = MODELS[model_name]["processor"]
|
59 |
+
|
60 |
+
pixel_values = torch.stack([processor(image) for image in [image_1, image_2]])
|
61 |
+
|
62 |
+
embeddings = model(pixel_values).pooler_output
|
|
|
63 |
embeddings /= embeddings.norm(p=2, dim=-1, keepdim=True)
|
64 |
|
65 |
+
similarity = (embeddings[0] @ embeddings[1].T).item()
|
66 |
+
|
67 |
return similarity
|
68 |
+
|
69 |
|
70 |
with gr.Blocks() as demo:
|
71 |
with gr.Row():
|
72 |
with gr.Column():
|
73 |
+
image_1 = gr.Image(label="Image 1", type="pil")
|
74 |
+
image_2 = gr.Image(label="Image 2", type="pil")
|
75 |
|
76 |
+
model_name = gr.Dropdown(
|
77 |
+
label="Model",
|
78 |
+
choices=list(MODELS.keys()),
|
79 |
+
value=list(MODELS.keys())[0],
|
80 |
+
)
|
81 |
submit_btn = gr.Button("Submit", variant="primary")
|
82 |
+
|
83 |
with gr.Column():
|
84 |
+
similarity = gr.Label(label="Similarity")
|
85 |
|
86 |
gr.on(
|
87 |
triggers=[submit_btn.click],
|
|
|
91 |
image_1,
|
92 |
image_2,
|
93 |
],
|
94 |
+
outputs=[similarity],
|
95 |
)
|
96 |
|
97 |
if __name__ == "__main__":
|
pyproject.toml
ADDED
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
[project]
|
2 |
+
name = "style-demo"
|
3 |
+
version = "0.1.0"
|
4 |
+
description = "Add your description here"
|
5 |
+
readme = "README.md"
|
6 |
+
requires-python = ">=3.11"
|
7 |
+
dependencies = [
|
8 |
+
"hf-xet>=1.0.3",
|
9 |
+
"safetensors>=0.5.3",
|
10 |
+
"timm>=1.0.15",
|
11 |
+
"torch>=2.6.0",
|
12 |
+
"torchvision>=0.21.0",
|
13 |
+
"transformers>=4.51.2",
|
14 |
+
]
|
15 |
+
|
16 |
+
[dependency-groups]
|
17 |
+
dev = [
|
18 |
+
"gradio>=5.25.0",
|
19 |
+
"ruff>=0.11.5",
|
20 |
+
]
|
requirements.txt
CHANGED
@@ -1,4 +1,5 @@
|
|
1 |
torch
|
|
|
2 |
transformers
|
3 |
timm
|
4 |
safetensors
|
|
|
1 |
torch
|
2 |
+
torchvision
|
3 |
transformers
|
4 |
timm
|
5 |
safetensors
|
uv.lock
ADDED
The diff for this file is too large to render.
See raw diff
|
|