我正在关注以 .h5 格式提供特定模型的教程。当然，我可以model.summary()在加载后调用这个模型load_model()，但是输出看起来像这样：

Layer (type)                 Output Shape              Param #   
=================================================================
conv1d_1 (Conv1D)            (None, 400, 32)           1568      
_________________________________________________________________
batch_normalization_1 (Batch (None, 400, 32)           128       
_________________________________________________________________
activation_1 (Activation)    (None, 400, 32)           0         
_________________________________________________________________
max_pooling1d_1 (MaxPooling1 (None, 100, 32)           0         
_________________________________________________________________
conv1d_2 (Conv1D)            (None, 100, 64)           32832     
_________________________________________________________________
batch_normalization_2 (Batch (None, 100, 64)           256       
_________________________________________________________________
activation_2 (Activation)    (None, 100, 64)           0         
_________________________________________________________________
max_pooling1d_2 (MaxPooling1 (None, 25, 64)            0         
_________________________________________________________________
conv1d_3 (Conv1D)            (None, 25, 128)           131200    
_________________________________________________________________
batch_normalization_3 (Batch (None, 25, 128)           512       
_________________________________________________________________
activation_3 (Activation)    (None, 25, 128)           0         
_________________________________________________________________
max_pooling1d_3 (MaxPooling1 (None, 6, 128)            0         
_________________________________________________________________
flatten_1 (Flatten)          (None, 768)               0         
_________________________________________________________________
dense_1 (Dense)              (None, 80)                61520     
_________________________________________________________________
batch_normalization_4 (Batch (None, 80)                320       
_________________________________________________________________
activation_4 (Activation)    (None, 80)                0         
_________________________________________________________________
dense_2 (Dense)              (None, 80)                6480      
_________________________________________________________________
batch_normalization_5 (Batch (None, 80)                320       
_________________________________________________________________
activation_5 (Activation)    (None, 80)                0         
_________________________________________________________________
dense_3 (Dense)              (None, 2)                 162       
_________________________________________________________________
activation_6 (Activation)    (None, 2)                 0         

我的问题：有没有办法找出使用了哪个激活函数？例如使用现有的 keras 方法的 relu、softmax、sigmoid 等？