Skip to content

[fix bug]激活函数不能使用relu，否则参数都为0，预测也全都是0 #3

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
zhudaoruyi wants to merge 1 commit into
base: master
Choose a base branch
from
Open

[fix bug]激活函数不能使用relu，否则参数都为0，预测也全都是0 #3

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
10 changes: 5 additions & 5 deletions README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -10,11 +10,11 @@ A Python 3 and Keras 2 implementation of MSCNN for people countingand provide tr
## MSCNN and MSB architectures
**MSCNN**

![MSCNN](/images/mscnn.png)
![MSCNN](images/mscnn.png)

**MSB**

![MSB](/images/msb.png)
![MSB](images/msb.png)

## Experiment

Expand All @@ -24,7 +24,7 @@ A Python 3 and Keras 2 implementation of MSCNN for people countingand provide tr

Generate density_map from data:

![density map](/images/density_map.png)
![density map](images/density_map.png)

**train**

Expand All @@ -38,7 +38,7 @@ python train.py --size 224 --batch 16 --epochs 100
real count:30
pred count:27
```
![res](/images/res1.png)
![res](images/res1.png)

## Reference

Expand All @@ -52,4 +52,4 @@ pred count:27

## Copyright
See [LICENSE](LICENSE) for details.

10 changes: 5 additions & 5 deletions model.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -16,7 +16,7 @@ def MSB(filters):
Returns:
f: function, layer func.
"""
params = {'activation': 'relu', 'padding': 'same',
params = {'activation': 'elu', 'padding': 'same',
'kernel_regularizer': l2(5e-4)}

def f(x):
Expand All @@ -26,7 +26,7 @@ def f(x):
x4 = Conv2D(filters, 3, **params)(x)
x = concatenate([x1, x2, x3, x4])
x = BatchNormalization()(x)
x = Activation('relu')(x)
x = Activation('elu')(x)

return x
return f
Expand All @@ -43,16 +43,16 @@ def MSCNN(input_shape):
"""
inputs = Input(shape=input_shape)

x = Conv2D(64, 9, activation='relu', padding='same')(inputs)
x = Conv2D(64, 9, activation='elu', padding='same')(inputs)
x = MSB(4 * 16)(x)
x = MaxPooling2D()(x)
x = MSB(4 * 32)(x)
x = MSB(4 * 32)(x)
x = MaxPooling2D()(x)
x = MSB(3 * 64)(x)
x = MSB(3 * 64)(x)
x = Conv2D(1000, 1, activation='relu', kernel_regularizer=l2(5e-4))(x)
x = Conv2D(1, 1, activation='relu')(x)
x = Conv2D(1000, 1, activation='elu', kernel_regularizer=l2(5e-4))(x)
x = Conv2D(1, 1, activation='elu')(x)

model = Model(inputs=inputs, outputs=x)

Expand Down
2 changes: 1 addition & 1 deletion test.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -38,5 +38,5 @@ def eva_regress(y_true, y_pred):
dmap = model.predict(img)[0][:, :, 0]
dmap = cv2.GaussianBlur(dmap, (15, 15), 0)

visualization(img[0], dmap)
visualization(img[0], np.expand_dims(dmap, axis=-1))
print('count:', int(np.sum(dmap)))
10 changes: 5 additions & 5 deletions train.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -53,13 +53,13 @@ def train(batch, epochs, size):
lr = ReduceLROnPlateau(monitor='loss', min_lr=1e-7)

indices = list(range(1500))
train, test = train_test_split(indices, test_size=0.25)
train_ids, test_ids = train_test_split(indices, test_size=0.25)

hist = model.fit_generator(
generator(train, batch, size),
validation_data=generator(test, batch, size),
steps_per_epoch=len(train) // batch,
validation_steps=len(test) // batch,
generator(train_ids, batch, size),
validation_data=generator(test_ids, batch, size),
steps_per_epoch=len(train_ids) // batch,
validation_steps=len(test_ids) // batch,
epochs=epochs,
callbacks=[lr])

Expand Down