Uriel Singer

> Hi @urielsinger, I tried to replicate your issue but I wasn't able to run to the same issue as you did. > > I used this command `wandb sync...

Notice that line 85-87 checks that all nodes have the same attributes. Deleting these 4 spaces deletes this check. If you are comfortable with this - it is possible. Line...

Hey, This is a beta version of the code. Working on pushing a final version.

In order to use the `node_classification` task, you should give the graph (`graph_nx2`) a node-attribute called `label` which holds the nodes labels. For further explanation, please refer to: https://networkx.org/documentation/stable/reference/generated/networkx.classes.function.set_node_attributes.html

You probably have a bug, I think it is best to debug.

Hi, I would suggest changing the model to return the LSTM sequence prediction and then choose the desired timestamp you want to predict on. technically, in the following line return_sequences=True:...

Hey holowrd, In the [paper](https://www.ijcai.org/proceedings/2019/0640.pdf) you can find on page 5 the links to all datasets. Uriel

Please check out the last two functions in `https://github.com/urielsinger/familyGan/blob/master/familyGan/pipeline.py`

Those two functions show the full pipe from raw images or files. Use the relevant parts of those functions as you wish.

Check your '_keras_' version. In version 1.2.2 it's '_nb_epoch_' while in later versions it's '_epochs_'.