由于contrib.rnn.ConvLSTMCell中对于ConvLSTMCell的实现本没有基于原作者的所引用的带有 "peepholes connection"的LSTM。因此，这里就照着葫芦画瓢，直接在原来的contrib.rnn.ConvLSTMCell的call()实现中上添加了peepholes这一步。
添加的代码为：
kernel_shape = cell.shape.as_list()[-3:]
w_ci = vs.get_variable(
"w_ci", kernel_shape, inputs.dtype)
w_cf = vs.get_variable(
"w_cf", kernel_shape, inputs.dtype)
w_co = vs.get_variable(
"w_co", kernel_shape, inputs.dtype)
new_cell = math_ops.sigmoid(forget_gate + self._forget_bias + w_cf * cell) * cell
new_cell += math_ops.sigmoid(input_gate + w_ci * cell) * math_ops.tanh(new_input)
output = math_ops.tanh(new_cell) * math_ops.sigmoid(output_gate + w_co * new_cell)引用时，将 ConvLSTM中的BasicConvLSTM导入即可：
from ConvLSTM import BasicConvLSTM用法同ConvLSTMCell一模一样！