ncnn-webassembly-nanodet
ncnn-webassembly-nanodet copied to clipboard
nihui
Deploy nanodet, the super fast and lightweight object detection, in your web browser with ncnn and webassembly
RuntimeError: Out of bounds memory access (evaluating '_nanodet_ncnn(dst, canvas.width, canvas.height)') /ai/ncnn-webassembly-nanodet/index.html:207:22 .wasm-function[339]@[wasm code] .wasm-function[364]@[wasm code] .wasm-function[363]@[wasm code] .wasm-function[17]@[wasm code] .wasm-function[28]@[wasm code] wasm-stub@[wasm code] 28@[native code] ncnn_nanodet@https://myoss.wmdigit.com/ai/ncnn-webassembly-nanodet/index.html:207:22 sFilter@https://myoss.wmdigit.com/ai/ncnn-webassembly-nanodet/index.html:222:21 iOS微信小程序webview 加载网页报错，无法识别出结果
Hi, wanted to ask if you have experienced a big drop in model accuracy when running inference in browser? The performance dropped quite significantly as compared to running the same...
请问其他版本行吗？我尝试了3.1.25，同样的程序，paddle-ocr v3在桌面上运行愉快，但在浏览器上的输出结果不对(detector的部分)。因为Qt6.5.2指定要用emscripten 3.1.25的关系，可能没办法更换。nanodet-plus-m的话运行起来倒是很顺利，至少肉眼看不出差别 编译的指令是参考文档的： ``` mkdir -p build-simd-threads cd build-simd-threads cmake -DCMAKE_TOOLCHAIN_FILE=$EMSDK/upstream/emscripten/cmake/Modules/Platform/Emscripten.cmake \ -DNCNN_THREADS=ON -DNCNN_OPENMP=ON -DNCNN_SIMPLEOMP=ON -DNCNN_SIMPLEOCV=ON -DNCNN_RUNTIME_CPU=OFF -DNCNN_SSE2=ON -DNCNN_AVX2=OFF -DNCNN_AVX=OFF \ -DNCNN_BUILD_TOOLS=OFF -DNCNN_BUILD_EXAMPLES=OFF -DNCNN_BUILD_BENCHMARK=OFF .. cmake --build . -j...
when i build it ,error occured ~/code/ncnn/ncnn-webassembly-nanodet-master/build$ cmake -DCMAKE_TOOLCHAIN_FILE=$EMSDK/upstream/emscripten/cmake/Modules/Platform/Emscripten.cmake -DWASM_FEATURE=basic .. CMake Error at /home/inpg/code/ncnn/emsdk/upstream/emscripten/cmake/Modules/Platform/Emscripten.cmake:131 (message): Failed to fetch compiler version information with command "'/home/inpg/code/ncnn/emsdk/upstream/emscripten/emcc' -v"! Process returned with error...
