DLTensor

mnn加载剪枝模型错误

1

剪枝后进行mnn转换，报错 Error for compute convolution shape, need channel = 62, input channel = 64 Error for compute convolution shape, need channel = 62, input channel = 64 Error for compute...

AttributeError: 'Image' object has no attribute 'shape'

1

Dear author，the environment is ready when I tests the demo, but I cannot train my data. The problem is occur when I run it. Can you give me some advises...

使用此工程非剪枝跑出来的精度低于使用darknet跑出来的模型精度； 分别使用如下训练方式： 1. 使用readme的正常训练方式进行训练 2. 将剪枝部分的代码注释掉训练 两种方式得到的模型精度，都比darknet训练的模型精度要低

No module named 'nvcv'

1

I have installed "nvcv_python-0.2.0_alpha-cp38-cp38-linux_x86_64.whl" in environment (python3.8.3 / cuda11.0 / ubuntu18.04). and then I code "import nvcv", but this message "no module named 'nvcv'" posts

inference costs 29ms 1080TI

2

I followed your README.md, modified Makefile to open CUDA CUDNN and GPU, and then run "make" I used ./Yolo-Fastest/VOC/yolo-fastest.weights to run image_yolov3.sh, it costs 29ms on 1080TI GPU. I do...