nihui
nihui
针对onnx模型转换的各种问题,推荐使用最新的pnnx工具转换到ncnn In view of various problems in onnx model conversion, it is recommended to use the latest pnnx tool to convert your model to ncnn ```shell pip install pnnx pnnx...
针对onnx模型转换的各种问题,推荐使用最新的pnnx工具转换到ncnn In view of various problems in onnx model conversion, it is recommended to use the latest pnnx tool to convert your model to ncnn ```shell pip install pnnx pnnx...
hi, ncnn yolo11 examples are on board https://github.com/Tencent/ncnn/tree/master/examples https://github.com/nihui/ncnn-android-yolo11 model conversion guide (zh) https://zhuanlan.zhihu.com/p/1903414797195781701
will be fixed in the next release
closed for https://github.com/Tencent/ncnn/pull/6233
test_shufflechannel did not go wrong, and I did not find any memory anomalies using valgrind. When I reviewed this issue again, I thought that only the last load exceeded some...
hi, yolov8 examples are updated with full support for detection, segmentation, classification, pose estimation and obb https://github.com/Tencent/ncnn/tree/master/examples android demo https://github.com/nihui/ncnn-android-yolov8 detailed instruction (zh) https://zhuanlan.zhihu.com/p/16030630352
是的 s3 有硬件浮点所以不量化也很快,量化后带来的额外 quant/dequant 操作反而会增加耗时
可能是手机性能较弱,尝试在加载模型前设置 `net.opt.num_threads` 减少cpu线程数,让出一些资源
https://github.com/Tencent/ncnn/wiki/FAQ-ncnn-produce-wrong-result#make-data-contiguous