bestpower

Android更新20220701与20220721两个大版本库后编译的JNI库在应用加载时报错

3

## error log | 日志或报错信息 ```txt E/art ( 9849): dlopen("/data/app/packagename-1/lib/arm/libname.so", RTLD_LAZY) failed: dlopen failed: cannot locate symbol "_ZNSt13runtime_errorD1Ev" referenced by "libname.so"... D/AndroidRuntime( 9849): Shutting down VM E/AndroidRuntime( 9849): FATAL EXCEPTION:...

最新更新编译出的benchmark程序测试存在性能负优化问题

3

## 日志或报错信息 旧版本测试结果： 816x32 ``` Convolution Conv_0 0.76ms | [816, 32, 3 *1] -> [408, 16, 2 *4] kernel: 3 x 3 stride: 2 x 2 HardSwish Div_0 0.36ms |...

相同输入的情况下onnx模型与其转换后的ncnn模型推理结果存在偏差

4

## 日志或报错信息 相同输入的情况下onnx模型与其转换后的ncnn模型推理结果数据中在不接近0或1的情况下存在较大偏差   ## 编译/运行环境 onnx推理环境：windows 10/Intel Core i7 9700/python==3.6/onnxruntime==1.4.0 ncnn推理环境：Android 10.0/展锐 SC9863A/8x ARM Cortex-A55 @1600MHz/ARMv8-A ncnn库版本commitid:6b2495cc243f2d8e829523b700f32db1f5d50f78 ## 复现步骤 | 再現方法 1.编写onnx模型推理代码 ```python # onnx模型推理关键代码 onnx_session = onnxruntime.InferenceSession(test_model.onnx)...

我之前提的问题仍未解决：使用修改Benchmark代码中的precision配置为PRECISION_LOW后推理模型出错 #1345

2

1345该问题是我之前提的，贵司只解决了一半便关闭了该问题，前一段时间因为在忙其他事情，未持续关注该问题，最近我这边拉最新的代码复测该问题仍存在，麻烦再看一下？详情请查看之前的问题记录。

Some error annotated landmarks in this Datasets

3

Hi autor, thank you for your sharing this dataset, it is very helpful in our face AI development, but I have found some errors(some annotated face points coordinates is not...

com.facebook.jni.CppException: view size is not compatible with input tensor's size and stride (at least one dimension spans across two contiguous subspaces). Use .reshape(...) instead.

1

In android app, my model output array as float[][], but can not get the output tensor as codes: ```java outputTensor = module.forward(IValue.from(inputTensor)).toTensor(); ``` How to solve please?

20220812最新代码下编译ncnn so库替换原有库后模型推理报错

5

## error log | 日志或报错信息 | ログ ```txt 08-22 11:03:52.112 6946 6990 F libc : Fatal signal 11 (SIGSEGV), code 1, fault addr 0xf300 in tid 6990 (Ocr_Detect_Work), pid 6946...

How to test a self trained model?

10

According to the readme, I have trained a model and got a new model file *.pt, but in this project I didn't find the codes to test the model. I...

下拉代码导入Android Studio无法build

1

一直提示说缺少libandroid_ceres.so库，请问你这个工程是不是缺了什么文件，代码上传有遗漏