youkee

> Try this cfg file: [yolov7_so.zip](https://github.com/WongKinYiu/yolov7/files/9286658/yolov7_so.zip) > > I used > > ``` > [-1, 1, nn.Upsample, [None, 4, 'nearest']], > [11, 1, Conv, [128, 1, 1]], # route backbone...

After I cleaned the cache file in the dataset file and train works.

Hi, I compile the tf_embedded mtcnn code on Ubuntu14.04 with Tensorflow1.4.0, I got same issue when run the binary file, but standalone mtcnn code works well. As to standalone mtcnn...

When I export the official model [yolov7](https://github.com/WongKinYiu/yolov7/releases/download/v0.1/yolov7.pt) to onnx with export.py from **u5** git brach, then use opencvdnn to do inference job. It works well. But when I train the...

> Please check this issue, and someone export the onnx working with opencvDNN and onnxruntime in C++ or python. https://github.com/WongKinYiu/yolov7/issues/145

> opencv dnn success run custom yolov7 > > 1 reparam trained best.pt https://github.com/WongKinYiu/yolov7/blob/main/tools/reparameterization.ipynb > > `nc = 1 #change with your nc` `model = Model('cfg/deploy/yolov7.yaml', ch=3, nc=nc).to(device)` `for i...

The preprocess and the postprocess part for own trained model are different the inference demo code, and you have to update the demo code according to train.py

从huggingface仓库看： 此版本为学术demo版，基于姜子牙[Ziya-LLaMA-13B-v1](https://huggingface.co/IDEA-CCNL/Ziya-LLaMA-13B-v1)训练而来(LLaMA权重的许可证限制，我们无法直接发布完整的模型权重，用户需自行合并) 就是说，需要模型合并