libofei2004
libofei2004
I met the same error in android sample object_detection.
@PaulTR Do you know how to set number of threads?
@PaulTR do you check on how to set number of threads?
@kuaashish ``` /* * Copyright 2022 The TensorFlow Authors. All Rights Reserved. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file...
``` /* * Copyright 2022 The TensorFlow Authors. All Rights Reserved. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except...
@kuaashish Could you see my codes?
@freedomtan thank you. I don't understand 1, how to check if it is really accelerated by NNAPI? My model has been quantized. My main code is: ``` /* * Copyright...
@Edge-cat [你好，想请教一下，我用你的方法改了相关代码，改了C2F和Detect层，ncnn-android-yolov8中也做了yolo.cpp&yolov8ncnn.cpp的修改，基于官方的yolov8n.pt转成yolov8n.bin和yolov8n.param，但运行后还是闪退，不知是哪里没做对
@dosubot 我的模型是官方下载的yolov8n.pt，以它为基础进行的训练，那么anchor数应该是多少？yolov8n原本的识别类别数是80，我只训练了一个类别，是否训练后的best.pt类别数就变成了1？
@dosu 我的问题就是不知道best_pnnx.py 脚本应该怎么改，看你的解释似乎应该是修改 view(1, 6 * anchor数, -1)，如果 anchor 数是1，就应该是view(1, 6 , -1)吗？把view(1, 144, -1)改成view(1, 6 , -1)？