Update app.py
Browse files
app.py
CHANGED
@@ -29,7 +29,7 @@ model.to(device=args.get("device"))
|
|
29 |
torch.set_grad_enabled(False)
|
30 |
|
31 |
|
32 |
-
def predict(image)->dict:
|
33 |
global model, args, normtransform
|
34 |
|
35 |
prediction_tensor = torch.zeros([1, len(args.get("dxlabels"))]).to(device=args.get("device"))
|
@@ -40,8 +40,8 @@ def predict(image)->dict:
|
|
40 |
aug_combos = [x for x in itertools.product(target_sizes, hflips, rotations, crops)]
|
41 |
|
42 |
# Load image
|
43 |
-
|
44 |
-
img =
|
45 |
|
46 |
# Predict with Test-time augmentation
|
47 |
for (target_size, hflip, rotation, crop) in aug_combos:
|
@@ -73,7 +73,7 @@ If you have a skin change in question, seek contact to your physician.'''
|
|
73 |
|
74 |
gr.Interface(
|
75 |
predict,
|
76 |
-
inputs=gr.Image(label="Upload a dermatoscopic image", type="
|
77 |
outputs=gr.Label(num_top_classes=len(args.get("dxlabels"))),
|
78 |
title="Dermatoscopic classification",
|
79 |
description=description,
|
|
|
29 |
torch.set_grad_enabled(False)
|
30 |
|
31 |
|
32 |
+
def predict(image: str)->dict:
|
33 |
global model, args, normtransform
|
34 |
|
35 |
prediction_tensor = torch.zeros([1, len(args.get("dxlabels"))]).to(device=args.get("device"))
|
|
|
40 |
aug_combos = [x for x in itertools.product(target_sizes, hflips, rotations, crops)]
|
41 |
|
42 |
# Load image
|
43 |
+
img = Image.open(image)
|
44 |
+
img = img.convert('RGB')
|
45 |
|
46 |
# Predict with Test-time augmentation
|
47 |
for (target_size, hflip, rotation, crop) in aug_combos:
|
|
|
73 |
|
74 |
gr.Interface(
|
75 |
predict,
|
76 |
+
inputs=gr.Image(label="Upload a dermatoscopic image", type="filepath"),
|
77 |
outputs=gr.Label(num_top_classes=len(args.get("dxlabels"))),
|
78 |
title="Dermatoscopic classification",
|
79 |
description=description,
|