Page History

...

• Solution (a): its advantage is that it's easy to have an efficient implementation and easy to deploy the model implemented with control flow operators. However, it's very difficult to implement and debug a model by using this interface.
• Solution (b): it's very intuitive to implement this a model. However, it's difficult to achieve efficiency and difficult to export the model for deployment. When implementing a dynamic model with the imperative interface, the frameworks need to With this approach, a DL framework can trace operators and construct a dynamic computation graph on the fly (for auto-differentiation and performance optimization. However, dynamic graph construction can cause significant overhead and it's difficult to optimize). The other problem is that a dynamic computation graph doesn't represent the neural network model, and thus can be exported for deployment.

...