Table 2. GRU sequence model summary for layers, output shape and number of parameters in each layer (keras summary function).

Layer (type)	Output shape	Number of parameters
input_2 (Input Layer)	[(None, 107, 8)]	0
tf_op_layer_strided_slice_2	[(None, 107, 3)]	0
embedding (Embedding)	(None, 107, 3, 100)	1,400
tf_op_layer_Reshape (TensorFlow)	[(None, 107, 300)]	0
tf_op_layer_strided_slice_1(TF)	[(None, 107, 3)]	0
concatenate (Concatenate)	(None, 107, 305)	0
bidirectional (Bidirectional)	(None, 107, 512)	861,696
bidirectional_1 (Bidirectional)	(None, 107, 512)	1,182,720
bidirectional_2 (Bidirectional)	(None, 107, 512)	1,182,720
tf_op_layer_strided_slice_4	[(None, 68, 512)]	0
dense (Dense)	(None, 68, 5)	2,565
Total params: 3,231,101
Trainable params: 3,231,101
Non-trainable params: 0