Aymeric Damien

What tensorflow version are you using?

Yes, it is the correct way, you simply need to run: ``` python # Before training, run this: tflearn.is_training(True, session=sess) # Before testing, run this: tflearn.is_training(False, session=sess) ```

So you should do something like that: ``` python # Training tflearn.is_training(True, session=sess) sess.run([train_op]) # Testing/Predicting tflearn.is_training(False, session=sess) sess.run([accuracy_op]) ```

You can have a check to `tflearn.layers.core.time_distributed`

You can add a parameter like this: ```python from tflearn.objectives import weighted_crossentropy tflearn.regression(net, optimizer='adam', learning_rate=0.01, loss=lambda x, y: weighted_crossentropy(x, y, weight=...)) ```

Hi, that is a good idea! I will take a look if I get some time, in the meanwhile, I mark this thread as "contributions welcome", if anyone is interested!