Spaces:
Running
Running
switched to pipeline
Browse files- app.py +7 -60
- resnetinceptionv1_epoch_32.pth +0 -3
app.py
CHANGED
@@ -10,29 +10,11 @@ import cv2
|
|
10 |
from pytorch_grad_cam import GradCAM
|
11 |
from pytorch_grad_cam.utils.model_targets import ClassifierOutputTarget
|
12 |
from pytorch_grad_cam.utils.image import show_cam_on_image
|
13 |
-
|
14 |
with zipfile.ZipFile("examples.zip","r") as zip_ref:
|
15 |
zip_ref.extractall(".")
|
16 |
|
17 |
-
|
18 |
-
|
19 |
-
mtcnn = MTCNN(
|
20 |
-
select_largest=False,
|
21 |
-
post_process=False,
|
22 |
-
device=DEVICE
|
23 |
-
).to(DEVICE).eval()
|
24 |
-
|
25 |
-
model = InceptionResnetV1(
|
26 |
-
pretrained="vggface2",
|
27 |
-
classify=True,
|
28 |
-
num_classes=1,
|
29 |
-
device=DEVICE
|
30 |
-
)
|
31 |
-
|
32 |
-
checkpoint = torch.load("resnetinceptionv1_epoch_32.pth", map_location=torch.device('cpu'))
|
33 |
-
model.load_state_dict(checkpoint['model_state_dict'])
|
34 |
-
model.to(DEVICE)
|
35 |
-
model.eval()
|
36 |
|
37 |
EXAMPLES_FOLDER = 'examples'
|
38 |
examples_names = os.listdir(EXAMPLES_FOLDER)
|
@@ -48,55 +30,20 @@ for example_name in examples_names:
|
|
48 |
np.random.shuffle(examples) # shuffle
|
49 |
|
50 |
def predict(input_image:Image.Image, true_label:str):
|
51 |
-
|
52 |
-
|
53 |
-
if face is None:
|
54 |
-
raise Exception('No face detected')
|
55 |
-
face = face.unsqueeze(0) # add the batch dimension
|
56 |
-
face = F.interpolate(face, size=(256, 256), mode='bilinear', align_corners=False)
|
57 |
-
|
58 |
-
# convert the face into a numpy array to be able to plot it
|
59 |
-
prev_face = face.squeeze(0).permute(1, 2, 0).cpu().detach().int().numpy()
|
60 |
-
prev_face = prev_face.astype('uint8')
|
61 |
-
|
62 |
-
face = face.to(DEVICE)
|
63 |
-
face = face.to(torch.float32)
|
64 |
-
face = face / 255.0
|
65 |
-
face_image_to_plot = face.squeeze(0).permute(1, 2, 0).cpu().detach().int().numpy()
|
66 |
-
|
67 |
-
target_layers=[model.block8.branch1[-1]]
|
68 |
-
use_cuda = True if torch.cuda.is_available() else False
|
69 |
-
cam = GradCAM(model=model, target_layers=target_layers, use_cuda=use_cuda)
|
70 |
-
targets = [ClassifierOutputTarget(0)]
|
71 |
-
|
72 |
-
grayscale_cam = cam(input_tensor=face, targets=targets, eigen_smooth=True)
|
73 |
-
grayscale_cam = grayscale_cam[0, :]
|
74 |
-
visualization = show_cam_on_image(face_image_to_plot, grayscale_cam, use_rgb=True)
|
75 |
-
face_with_mask = cv2.addWeighted(prev_face, 1, visualization, 0.5, 0)
|
76 |
-
|
77 |
-
with torch.no_grad():
|
78 |
-
output = torch.sigmoid(model(face).squeeze(0))
|
79 |
-
prediction = "real" if output.item() < 0.5 else "fake"
|
80 |
-
|
81 |
-
real_prediction = 1 - output.item()
|
82 |
-
fake_prediction = output.item()
|
83 |
-
|
84 |
-
confidences = {
|
85 |
-
'real': real_prediction,
|
86 |
-
'fake': fake_prediction
|
87 |
-
}
|
88 |
return confidences, true_label, face_with_mask
|
89 |
|
90 |
interface = gr.Interface(
|
91 |
fn=predict,
|
92 |
inputs=[
|
93 |
-
gr.
|
94 |
"text"
|
95 |
],
|
96 |
outputs=[
|
97 |
-
gr.
|
98 |
"text",
|
99 |
-
gr.
|
100 |
],
|
101 |
examples=[[examples[i]["path"], examples[i]["label"]] for i in range(10)]
|
102 |
).launch()
|
|
|
10 |
from pytorch_grad_cam import GradCAM
|
11 |
from pytorch_grad_cam.utils.model_targets import ClassifierOutputTarget
|
12 |
from pytorch_grad_cam.utils.image import show_cam_on_image
|
13 |
+
from transformers import pipeline
|
14 |
with zipfile.ZipFile("examples.zip","r") as zip_ref:
|
15 |
zip_ref.extractall(".")
|
16 |
|
17 |
+
pipe = pipeline(model="not-lain/deepfake",trust_remote_code=True)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
18 |
|
19 |
EXAMPLES_FOLDER = 'examples'
|
20 |
examples_names = os.listdir(EXAMPLES_FOLDER)
|
|
|
30 |
np.random.shuffle(examples) # shuffle
|
31 |
|
32 |
def predict(input_image:Image.Image, true_label:str):
|
33 |
+
out = pipe.predict(input_image)
|
34 |
+
confidences,face_with_mask = out["confidences"], out["face_with_mask"]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
35 |
return confidences, true_label, face_with_mask
|
36 |
|
37 |
interface = gr.Interface(
|
38 |
fn=predict,
|
39 |
inputs=[
|
40 |
+
gr.Image(label="Input Image", type="filepath"),
|
41 |
"text"
|
42 |
],
|
43 |
outputs=[
|
44 |
+
gr.Label(label="Class"),
|
45 |
"text",
|
46 |
+
gr.Image(label="Face with Explainability")
|
47 |
],
|
48 |
examples=[[examples[i]["path"], examples[i]["label"]] for i in range(10)]
|
49 |
).launch()
|
resnetinceptionv1_epoch_32.pth
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:794ebe83c6a7d7959c30c175030b4885e2b9fa175f1cc3e582236595d119f52b
|
3 |
-
size 282395989
|
|
|
|
|
|
|
|