-
出现的问题和#433《mpp解码后数据拷贝占用CPU,是否有更好方式进行拷贝?》类似,
目前使用硬件平台:rk3588
mpp版本:release
视频渲染软件使用 Qt5.15+OPENGL ES 3.2,将解码后的数据拷贝到GPU,利用OPENGL中的转换矩阵对YUV420SP数据转换并渲染
问题:
在从mpp的buffer中memcpy到OPENGL的PBO中时(数据量大小:192…
-
With rk3588 coming with 32G of memory, and with blender running, is their an 64bit ARM release coming please? The rk3588 may be a node for a farm?
-
编译内核文件时报下面的错误,我确实也没找到该变量的申明.
drivers/rknpu/rknpu_devfreq.c:237:25: error: ‘rockchip_opp_set_low_length’ undeclared here (not in a function)
237 | .set_soc_info = rockchip_opp_set_low_l…
elcky updated
3 weeks ago
-
RK3588板环境
系统版本:ubuntu 22.04
rknpu2-dev: 1.5.2
PC环境:
系统版本:ubuntu 22.04
python: 3.8
rknn-toolkit2: 1.5.2+b642f30c
问题描述:
1)YOLOv8-seg模型量化精度异常 2)rknn-toolkit2配置未生效
复现步骤:
模型:官方yolov8s-seg.p…
-
### Feature description
Hi, I'd love for the docker containers to be made available for aarch64. I have a capable RK3588 board with 32GB RAM and was able to build the containers manually, but having …
-
I slightly modified the code in examples/resnet/cpp/main to measure execution time. measurements showed different results on linux and android, with the first run out of 100 showing approximately th…
-
/home/linaro/rknn-toolkit2-master/rknn_toolkit_lite2/examples/inference_with_lite/rknn-cpp-Multithreading-main
模型名称: ./model/RK3588/yolov5s-640-640_rm_transpose_rk3588.rknn
线程数: 1
Loading mod…
-
How to set parameters when Yolov5 is converted to onnx? I can use this onnx (rknpu2/examples/rknn_yolov5_demo/convert_rknn_demo/yolov5/onnx_models/yolov5s_rm_transpose.onnx) to convert to rknn and com…
-
我尝试转化7B的量化模型都失败了,AWQ提示仅支持GPU,GPTQ则提示.float()不支持量化模型,请问你转化过量化模型吗,或者说是否无法转化量化模型?我看test中有一个loss.float(), 但是感觉也无从下手。谢谢!
-
由于项目需要单线程运行,对于YOLO-fastestv2,在amd平台的虚拟机和rk3588上单线程设置都是正常的,查看cpu占用一般只有一到两个核心高。但是配置在Fastestdet的用同样设置上对于虚拟机平台,1-12cpu都在30-40%,而对于rk3588,1-8cpu都在70%左右。
上述都是设置的单线程。使用
ncnn::Net net;
net.opt.num_…
Zi-x updated
2 months ago