Stéphane Charette

I also tried it on a different computer where Darknet is compiled to use CPU-only, and it seemed to work fine on that computer as well: 

Note I tested on a NVIDIA 2070, 3050, and 3090, and it seems to work fine for me. You'd have to provide some steps for me to reproduce the problem.

Can this be closed?

ONNX support is not completed but in development in the v5_onnx branch. Hope to have it done in time for the V5 release later in 2025.

Two different Windows users in the last few days on the Darknet/YOLO discord server have reported what I believe might be this error while running the following command: `.\vcpkg.exe install...

Apparently, users have reported the fix is this: https://github.com/microsoft/vcpkg/issues/34247#issuecomment-2227320165 The exact message was: > ok i found this issue somehow: https://github.com/microsoft/vcpkg/issues/34247 > > and ended up using this: https://github.com/abbodi1406/vcredist to...

So far I cannot replicate. I've tried both yolov4-tiny at 416x416 and yolov4 at 608x608. I'm currently in the middle of training the YOLOv4 one. You are using the up-to-date...

Yes, that is the dataset I downloaded as well. It is working for me. The difference I can see from the command you give above is I use DarkMark to...

If you diff the two config files, you'll see the learning rate is the big difference between the two.

> https://github.com/AlexeyAB/darknet?tab=readme-ov-file#how-to-train-with-multi-gpu > > > ## How to train with multi-GPU > > > > 1. Train it first on 1 GPU for like 1000 iterations: `darknet.exe detector train cfg/coco.data...