10 lines
275 B
Markdown
10 lines
275 B
Markdown
Usage of Hybrid Quantization
|
|
```
|
|
1. install RKNN-Toolkit
|
|
2. execute commands below:
|
|
a. python step1.py
|
|
b. according to the result of accuracy_analysis, select the operand with more accuracy loss.
|
|
c. modify ssd_mobilenet_v2.quantization.cfg
|
|
d. python step2.py
|
|
```
|