RKNN部署记录-2（RKNN python API安装，使用

查看Linux磁盘容量：

df -h

 查询NPU驱动版本：

 dmesg | grep -i rknpu 
# 或
 cat /sys/kernel/debug/rknpu/version 
# 或
 cat /sys/kernel/debug/rknpu/driver_version 
# 或
 cat /proc/debug/rknpu/driver_version

查询python版本并退出

python3

exit()

安装pip，numpy，ruamel.yaml，psutils

sudo apt-get update

apt-get install python3-pip

sudo apt-get install -y python3-numpy

sudo apt-get install ruamel.yaml

sudo apt-get install psutils

安装rknn-toolkit lite 2

# Python 3.10
 pip3 install rknn_toolkit_lite2-x.y.z-cp310-cp310-linux_aarch64.whl

python脚本运行板端推理

import time
import numpy as np
import platform
from rknnlite.api import RKNNLite
input_data = np.load('./test_input.npy')
rknn_lite = RKNNLite(verbose = False)
i = 0

print('--> Load RKNN model')
ret = rknn_lite.load_rknn('mobilenet_v1.rknn')
if ret != 0:
    print('Load RKNN model failed')
    exit(ret)
print('done')

print('--> Init runtime environment')
ret = rknn_lite.init_runtime()
if ret != 0:
   print('Init runtime environment failed')
   exit(ret)
print('done')

while i<100:
	start_time = time.time()
	# Inference
	print('--> Running model')
	outputs = rknn_lite.inference(inputs=[input_data])

	end_time = time.time()
	execution_time = end_time - start_time
	print(execution_time)
	i+=1

rknn_lite.release()
print('done')

运行时报错，

invalid rknn model version 6

将  .\rknpu2\runtime\Linux\librknn_api\aarch64\librknnrt.so 文件复制到Linux下 /usr/bin/目录下即可成功运行

还有一些遇到的问题见RKNN部署记录-1-CSDN博客