Try documentation again

class LayerSplitter(Layer):

  """

  Layer which takes a tensor from in_tensor[0].out_tensors at an index

  This is a special layer which only makes sense in the context of TensorGraph.

  It takes a single input layer which sets the class variable self.out_tensors

  to a list of tensors.

  Only layers which need to output multiple layers set and use the variable

  self.out_tensors.

  This is a utility for those special layers which set self.out_tensors

  to return a layer wrapping a specific tensor in in_layers[0].out_tensors

  """

  def __init__(self, output_num, **kwargs):

import tensorflow as tf

from deepchem.models import TensorGraph

from deepchem.models.tensorgraph.graph_layers import Combine_AP, Separate_AP, \

  WeaveLayer, WeaveGather, DTNNEmbedding, DTNNGather, DTNNStep, \

from deepchem.models.tensorgraph.graph_layers import WeaveLayer, WeaveGather, DTNNEmbedding, DTNNGather, DTNNStep, \

  DTNNExtract, DAGLayer, DAGGather, MessagePassing, SetGather

from deepchem.models.tensorgraph.layers import Feature, Conv1D, Dense, Flatten, Reshape, Squeeze, Transpose, \

  CombineMeanStd, Repeat, Gather, GRU, L2Loss, Concat, SoftMax, \