1

I just started to learn tensorlayer, an excellent high-level wrapper over tensorflow framework for deep learning. Does anyone know how to do parameter gridsearch to tune DNN? Is there anyway that I can use GridSearchCV from sklearn to tune network parameters from tensorlayer? Thanks a lot.

Jonathan DEKHTIAR
  • 3,456
  • 1
  • 21
  • 42
zesla
  • 11,155
  • 16
  • 82
  • 147
  • This is a fine question. Probably not very useful, but I have found a [distributed example of GridSearchCV with Spark](https://databricks.com/blog/2016/02/08/auto-scaling-scikit-learn-with-apache-spark.html). I didn't know about TensorLayer: Would you recommend it over Keras? – lucid_dreamer Jun 09 '17 at 22:44

1 Answers1

0

you can do something like this.

for .... (different hyper-parameter):
    with tf.Graph().as_default() as graph:  # clear all variables of TF
       tl.layers.clear_layers_name()         # clear all layer name of TL
       sess = tf.InteractiveSession()
       # define and train a new model here

ref: https://github.com/tensorlayer/tensorlayer/issues/214

zsdh
  • 126
  • 1
  • 7