fixed model shape
Browse files
app.py
CHANGED
@@ -142,15 +142,9 @@ def predict_and_analyze(model_name, num_channels, dim, image):
|
|
142 |
# im = f.readlines()
|
143 |
# image = np.frombuffer(image)
|
144 |
|
145 |
-
print(image)
|
146 |
-
print(type(image))
|
147 |
-
print(image.name)
|
148 |
-
|
149 |
image = np.load(image.name, allow_pickle=True)
|
150 |
|
151 |
-
image = image.reshape((num_channels, W, W))
|
152 |
-
print(image)
|
153 |
-
print(type(image))
|
154 |
|
155 |
# W = int(np.sqrt(image.shape[1]))
|
156 |
|
@@ -159,6 +153,8 @@ def predict_and_analyze(model_name, num_channels, dim, image):
|
|
159 |
if len(image.shape) != 4:
|
160 |
image = image[np.newaxis, :, :, :]
|
161 |
|
|
|
|
|
162 |
input_image = np.sum(image[0, :, :, :], axis=0)
|
163 |
|
164 |
model_name += '_%i' % (num_channels)
|
|
|
142 |
# im = f.readlines()
|
143 |
# image = np.frombuffer(image)
|
144 |
|
|
|
|
|
|
|
|
|
145 |
image = np.load(image.name, allow_pickle=True)
|
146 |
|
147 |
+
# image = image.reshape((num_channels, W, W))
|
|
|
|
|
148 |
|
149 |
# W = int(np.sqrt(image.shape[1]))
|
150 |
|
|
|
153 |
if len(image.shape) != 4:
|
154 |
image = image[np.newaxis, :, :, :]
|
155 |
|
156 |
+
assert image.shape == (1, num_channels, W, W), "Data is the wrong shape"
|
157 |
+
|
158 |
input_image = np.sum(image[0, :, :, :], axis=0)
|
159 |
|
160 |
model_name += '_%i' % (num_channels)
|