Kusunoki0130 comments

Results 3 comments of


                                            Kusunoki0130

运行模型输出推理正常，但是一直提示W RKNN:Output(output): size_with_stride larger than model origin size, if need run Outputoperator in NPU, please call rknn_create_memory using s ize _with_stride.

I also see this problem. You can try to use librknnrt.so 1.6.0 in their new repository. Just replace the old verison librknnrt.so in /usr/lib.

rknn_toolkit_lite2_1.5.0 推理yolov5S 出现错误

I also see this problem. You can try to use librknnrt.so 1.6.0 in their new repository. Just replace the old verison librknnrt.so in /usr/lib.

controlNet - convert sketch circuit to vector-engineered circuit,some question

Maybe Diffusion Model is hard to recognize circuit components as it to texts or charaters，when you try to reconstruct images with them. 个人猜测：电路元件比较像文字符号，diffusion 模型这一系列我感觉它看不懂文字符号，没有办法细致地重构图像里文字，同样也不能很好地理解电路元件的排列。