Spaces:
Sleeping
Sleeping
update
Browse files- app.py +44 -14
- requirements.txt +3 -1
app.py
CHANGED
@@ -3,22 +3,56 @@ import numpy as np
|
|
3 |
import gradio as gr
|
4 |
from lavis.models import load_model_and_preprocess
|
5 |
from PIL import Image
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
6 |
|
7 |
def process(input_image, prompt):
|
8 |
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
|
9 |
|
10 |
-
model, vis_processors, txt_processors = load_model_and_preprocess(name="
|
11 |
-
|
12 |
-
input_image = input_image.resize((256, 256), Image.LANCZOS)
|
13 |
image = vis_processors["eval"](input_image).unsqueeze(0).to(device)
|
14 |
text_input = txt_processors["eval"](prompt)
|
15 |
sample = {"image": image, "text_input": [text_input]}
|
16 |
|
17 |
-
|
18 |
-
|
19 |
-
|
20 |
-
|
21 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
22 |
|
23 |
return preds
|
24 |
|
@@ -29,11 +63,7 @@ if __name__ == '__main__':
|
|
29 |
input_image, prompt
|
30 |
]
|
31 |
outputs = "image"
|
32 |
-
input_size = (256, 256)
|
33 |
-
output_size = (256, 256)
|
34 |
iface = gr.Interface(fn=process,
|
35 |
inputs=ips,
|
36 |
-
outputs=outputs
|
37 |
-
|
38 |
-
output_size=output_size)
|
39 |
-
iface.launch()
|
|
|
3 |
import gradio as gr
|
4 |
from lavis.models import load_model_and_preprocess
|
5 |
from PIL import Image
|
6 |
+
import matplotlib.pyplot as plt
|
7 |
+
import matplotlib.cm as cm
|
8 |
+
import torchvision
|
9 |
+
|
10 |
+
def create_heatmap(activation_map):
|
11 |
+
# アクティベーションマップをnumpy配列に変換
|
12 |
+
activation_map_np = activation_map.squeeze().detach().cpu().numpy()
|
13 |
+
|
14 |
+
# アクティベーションマップの最小値と最大値を取得
|
15 |
+
min_value = np.min(activation_map_np)
|
16 |
+
max_value = np.max(activation_map_np)
|
17 |
+
|
18 |
+
# アクティベーションマップを0-1の範囲に正規化
|
19 |
+
normalized_map = (activation_map_np - min_value) / (max_value - min_value)
|
20 |
+
|
21 |
+
# 正規化されたアクティベーションマップをヒートマップに変換
|
22 |
+
heatmap = cm.jet(normalized_map)
|
23 |
+
|
24 |
+
# ヒートマップを [0, 255] の範囲にスケーリングし、uint8型に変換
|
25 |
+
heatmap = np.uint8(255 * heatmap)
|
26 |
+
|
27 |
+
return heatmap
|
28 |
|
29 |
def process(input_image, prompt):
|
30 |
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
|
31 |
|
32 |
+
model, vis_processors, txt_processors = load_model_and_preprocess(name="pnp_vqa", model_type="base", is_eval=True, device=device)
|
33 |
+
input_image = input_image.resize((256, 256))
|
|
|
34 |
image = vis_processors["eval"](input_image).unsqueeze(0).to(device)
|
35 |
text_input = txt_processors["eval"](prompt)
|
36 |
sample = {"image": image, "text_input": [text_input]}
|
37 |
|
38 |
+
output = model.forward_itm(samples=sample)
|
39 |
+
activation_map = output['gradcams'].reshape(24, 24)
|
40 |
+
|
41 |
+
relu = torch.nn.ReLU()
|
42 |
+
# ヒートマップを計算
|
43 |
+
heatmap = create_heatmap(activation_map)
|
44 |
+
|
45 |
+
heatmap = Image.fromarray(heatmap)
|
46 |
+
heatmap = torchvision.transforms.functional.to_tensor(heatmap)
|
47 |
+
heatmap = relu(heatmap)
|
48 |
+
heatmap = torchvision.transforms.functional.to_pil_image(heatmap)
|
49 |
+
heatmap = heatmap.resize((256, 256))
|
50 |
+
heatmap = np.array(heatmap)
|
51 |
+
heatmap = torch.sigmoid(torch.from_numpy(heatmap)).numpy()
|
52 |
+
preds = heatmap.reshape(256, 256, -1)
|
53 |
+
preds = Image.fromarray(preds.astype(np.uint8)).convert('L')
|
54 |
+
preds = np.array(preds)
|
55 |
+
preds = np.where(preds > 0.5, 255, 0)
|
56 |
|
57 |
return preds
|
58 |
|
|
|
63 |
input_image, prompt
|
64 |
]
|
65 |
outputs = "image"
|
|
|
|
|
66 |
iface = gr.Interface(fn=process,
|
67 |
inputs=ips,
|
68 |
+
outputs=outputs)
|
69 |
+
iface.launch()
|
|
|
|
requirements.txt
CHANGED
@@ -3,4 +3,6 @@ transformers
|
|
3 |
torch
|
4 |
pillow
|
5 |
salesforce-lavis==1.0.2
|
6 |
-
numpy
|
|
|
|
|
|
3 |
torch
|
4 |
pillow
|
5 |
salesforce-lavis==1.0.2
|
6 |
+
numpy
|
7 |
+
torchvision
|
8 |
+
matplotlib
|