Using the same list of tensors to represent different lists in keras models

[khatchad]

We should have a seperate list for each of these attributes.

ML/com.ibm.wala.cast.python.ml/data/tensorflow.xml

Lines 749 to 755 in 16acfbc

	<new def="x" class="Llist" />
	<call class="LRoot" name="read_data" descriptor="()LRoot;" type="virtual" arg0="arg0" def="xx" />
	<putfield class="LRoot" field="0" fieldType="LRoot" ref="x" value="xx" />
	<!-- https://www.tensorflow.org/guide/keras/transfer_learning#freezing_layers_understanding_the_trainable_attribute -->
	<putfield class="LRoot" field="trainable_weights" fieldType="LRoot" ref="arg0" value="x" />
	<putfield class="LRoot" field="weights" fieldType="LRoot" ref="arg0" value="x" />
	<putfield class="LRoot" field="non_trainable_weights" fieldType="LRoot" ref="arg0" value="x" />