search

intel-analytics / BigDL-Tutorials

Step-by-step Deep Leaning Tutorials on Apache Spark using BigDL
208 stars 123 forks source link

Why there is no bias term? #6

Closed cmusjtuliuyuan closed 7 years ago

cmusjtuliuyuan commented 7 years ago

There is no bias term in linear_regression.ipynb, deep_feed_forward_neural_network.ipynb, cnn.ipynob ... I don't understand why?

For example, in the BigDL tutorials: model.add(Linear(n_input, n_hidden_1).set_name('mlp_fc1')) model.add(ReLU()) When these two lines implemened in Tensorflow or Pytorch, it will look like: W = weight_variable([input_dim, output_dim]) b = bias_variable([output_dim]) logits=tf.nn.relu(tf.matmul(input_placeholder, W) + b)

cmusjtuliuyuan commented 7 years ago

I understand the bias term is included into the Linear