Learning Wizard
This section describes the intended audience, porting process of a TensorFlow 1.15-based model, and precautions for using this document.
Intended Audience
This document is intended for AI algorithm engineers and describes how to port network scripts developed using TensorFlow 1.15-based Python APIs to run on the NPU. The NPU supports porting of network scripts developed with three TensorFlow 1.15 APIs: Estimator, sess.run, and Keras.
To better understand this document, you should:
- Be familiar with the basic CANN architecture and features.
- Be familiar with TensorFlow 1.15 APIs.
- Possess knowledge on machine learning and deep learning, especially network training basics.
- Be proficient in Python programming.
Supported Products
Atlas 350 Accelerator Card
Precautions
- Before model porting to the NPU, prepare the training model developed based on TensorFlow 1.15 and the corresponding dataset. Ensure that it runs successfully on a GPU or CPU and meets the expected accuracy and performance requirements, and record the relevant accuracy and performance metrics for subsequent comparison on the NPU.
- The code snippets in this document are for reference only. Modify and adapt them accordingly before use.
Model Development Learning Map
The model development aims to port the original TensorFlow-based model to the AI processor and start training. The process is as follows:
