Document Description
Document Name |
Description |
Link |
|---|---|---|
Script Conversion Tool Instructions |
GPU-based training and online inference scripts cannot be directly migrated to the NPU. |
|
Profiling Tool Instructions |
Describes how to use the Profiling tool. The tool is used to analyze key performance bottlenecks in the training phase and provide performance optimization suggestions to achieve optimal product performance. |
|
Auto Tune Tool Instructions |
During network model training in a training scenario, the operator search tool can automatically optimize operators during network model compilation. |
|
Model Accuracy Analyzer Instructions |
Huawei supports the migration of the original network for training on the Ascend 910 AI Processor. As a result, the computation result of Huawei proprietary operators may be different from that of third-party operators (for example, from TensorFlow). The Model Accuracy Analyzer is designed to compare the computation results for developers to quickly resolve the operator accuracy issues. |
|
AICore Error Analyzer Instructions |
When you encounter AI Core errors in the training when executing an app project, you can use the AICore Error Analyzer to collect necessary information for quickly locating the AI Core errors. |
|
ADC Instructions |
Describes how to use the Ascend Debug Client (ADC) commands to transfer files to the host or device, set log levels, and detect the heartbeat between the host and the device. |