为何剪枝后的TensorFlow模型文件体积比初始模型更大？

我正在参考TensorFlow官方的剪枝示例探索神经网络剪枝，基于预训练模型的剪枝代码如下：

prune_low_magnitude = tfmot.sparsity.keras.prune_low_magnitude

# Compute end step to finish pruning after 2 epochs.
batch_size = 64
epochs = 3
validation_split = 0.1 # 10% of training set will be used for validation set.

num_images = 114 * (1 - validation_split)
end_step = np.ceil(num_images / batch_size).astype(np.int32) * epochs

# Define model for pruning.
pruning_params = {'pruning_schedule': tfmot.sparsity.keras.PolynomialDecay(initial_sparsity = 0.50,
                                                               final_sparsity = 0.80,
                                                               begin_step = 0,
                                                               end_step = end_step)
}

pruned_model = prune_low_magnitude(model, **pruning_params)

# `prune_low_magnitude` requires a recompile.
pruned_model.compile(optimizer = 'adam', loss = keras.losses.SparseCategoricalCrossentropy(from_logits = True), metrics = ['accuracy'])

logdir = tempfile.mkdtemp()

callbacks = [
  tfmot.sparsity.keras.UpdatePruningStep(),
  tfmot.sparsity.keras.PruningSummaries(log_dir = logdir),
]

pruned_model.fit(train_dataset, batch_size=batch_size, epochs=epochs, validation_data=valid_dataset, callbacks=callbacks)

通过pruned_model.evaluate(train_dataset, verbose=0)验证，剪枝模型准确率略有下降，符合预期，测试结果如下：

Baseline test accuracy: 0.9197102189064026
Pruned model test accuracy: 0.8976686000823975

我使用model.save()将初始模型和剪枝模型分别保存为.h5和.keras格式：

• 初始模型大小约为60.7-60.9MB
• 剪枝模型的.h5格式为85.4MB，.keras格式更是达到110MB

我在Keras文档中未找到保存剪枝模型时需指定优化的相关说明。

内容的提问来源于stack exchange，提问作者Barney_Dinosaur