Convert a Tensorflow model in SavedModel format (.pb file) saved with tf.saved_model.save to a Keras model (.h5 file)

Question

I'm trying to convert a Tensorflow object detection model that's saved in a SavedModel format (in a folder with .pb file, assets folder, and variables folder) to Keras.h5 for simpler inference. 但是，尝试对此问题的常见 StackOverflow 答案（例如this ）会导致ValueError: Unable to create a Keras model from this SavedModel. This SavedModel was created with tf.saved_model.save, and lacks the Keras metadata.Please save your Keras model by calling model.save or tf.keras.models.save_model. ValueError: Unable to create a Keras model from this SavedModel. This SavedModel was created with tf.saved_model.save, and lacks the Keras metadata.Please save your Keras model by calling model.save or tf.keras.models.save_model. .

model 使用 Tensorflow Object 检测 ZDB974238714CA8A1434A7CE1D08 进行训练和保存。 是否仍然可以将其转换为 Keras model？ 如果是这样，有什么建议吗？

Answer 1

在将 YoloV5 转换为 Tensorflow 格式以部署 model 时，我遇到了同样的问题。 您需要在 model 的导出命令中使用keras=True 。 这将在目录中生成 Keras_metadata.pb 文件。

[https://www.tensorflow.org/api_docs/python/tf/keras/models/save_model][1]

[1]：请参考此链接添加上述命令以保存 Tensorflow model。