In Caffe, the SGD solver has a momentum parameter (link). In TensorFlow, I see that tf.train.GradientDescentOptimizer does not have an explicit momentum parameter. However, I can see that there is tf.train.MomentumOptimizer optimizer. Is it the equivalent of Caffe SGD with momentum optimizer?
Yes it is. tf.train.MomentumOptimizer = SGD + momentum