AlexeyAB/darknet issues and pull requests

#8921 - Dear Author, I would also like to know why the backward propagation for max pooling does not use the cuDNN API.

Issue - State: open - Opened by zzk2021 4 days ago - 1 comment

#8920 - log

Issue - State: closed - Opened by huhtig 7 days ago

#8919 - rnn can't generate text

Issue - State: open - Opened by zzk2021 about 1 month ago

#8918 - AttributeError: module 'darknet' has no attribute 'POINTER'

Issue - State: open - Opened by psilonpneuma 3 months ago - 1 comment

#8917 - YOLOv3 with PaNet, SPP and scale for five heads.

Issue - State: open - Opened by teptgeo 3 months ago
Labels: Feature-request

#8916 - add -Wextra and remove a lot of warnings for C code

Pull Request - State: open - Opened by uecker 3 months ago - 4 comments

#8915 - add -Wextra and make code warning free

Pull Request - State: closed - Opened by uecker 3 months ago - 3 comments

#8914 - Public contribution: darknet-to-pytorch-onnx

Issue - State: closed - Opened by ltetrel 3 months ago

#8913 - batch size and subdivision

Issue - State: open - Opened by NhatTamNL 3 months ago - 2 comments

#8911 - Convert yolov4-darknet object-detection model to Segmentation Model

Issue - State: open - Opened by karishmathumu 4 months ago

#8910 - what smaller model size is the most effective?

Issue - State: open - Opened by NhatTamNL 4 months ago - 4 comments

#8908 - yolo-seg

Issue - State: open - Opened by zengwb-lx 4 months ago

#8907 - Problem with assembly in windows 10

Issue - State: open - Opened by Chikibabah 5 months ago - 1 comment

#8905 - YoloV2 with different backbone

Issue - State: open - Opened by Iongng198 5 months ago

#8904 - mAP is 0% and the chart goes up and down...

Issue - State: open - Opened by hechoel 5 months ago - 4 comments
Labels: Training issue

#8903 - Models trained in Linux run poorly on Windows and vice-versa.

Issue - State: open - Opened by acidtonic 5 months ago - 7 comments

#8902 - fix memleak

Pull Request - State: closed - Opened by cenit 5 months ago

#8901 - docker fixes

Pull Request - State: closed - Opened by cenit 5 months ago

#8900 - build.ps1 is arm64 compatible

Pull Request - State: closed - Opened by cenit 5 months ago

#8899 - How do the parameters "ignore_thresh" and "iou_thresh" contribute to the training of a yolo detector model ?

Issue - State: open - Opened by riririririi 6 months ago

#8898 - Cuda stream support

Issue - State: open - Opened by dannyjpereira 6 months ago

#8897 - Dockerfile contains typos & references removed image

Issue - State: closed - Opened by patrickjdarrow 6 months ago

#8896 - YOLOv7 minimum in training loss

Issue - State: open - Opened by cschnu 6 months ago
Labels: Training issue

#8895 - make darknet directory

Issue - State: closed - Opened by AkumaBiNoTori 6 months ago - 1 comment

#8894 - Cannot download default weights file for tiny-yolo-voc.

Issue - State: open - Opened by chingkhei-th 6 months ago - 2 comments

#8893 - Fixed Missing 'nbiases' Item in YOLO Layer

Pull Request - State: closed - Opened by K9714 6 months ago - 1 comment

#8892 - l.iou_loss is alaways null in yolo_layer.c ,functon process_batch

Issue - State: closed - Opened by zzk2021 6 months ago - 1 comment

#8891 - hello,in box.c function dx_box_iou,why add twice grad when Iw<=0||Ih<=0

Issue - State: closed - Opened by zzk2021 6 months ago - 1 comment

#8890 - how to convert pt to weights in yolov7-tiny?

Issue - State: open - Opened by kdeltak 6 months ago

#8889 - I compiled the darknet. But when train, it has a lot beep sound. Can i turn it off?

Issue - State: closed - Opened by avkwok 7 months ago - 3 comments

#8888 - Update detector.c

Pull Request - State: closed - Opened by coleaskey32 7 months ago - 1 comment

#8887 - YOLOv9 is better than any convolution or transformer based object detectors

Issue - State: open - Opened by AlexeyAB 7 months ago - 6 comments

#8886 - Does the SAM-Mish activation function have a formula?

Issue - State: closed - Opened by ridvanozdemir 8 months ago

#8885 - Free Layer has a Memory Leak

Issue - State: closed - Opened by bernardobreder 8 months ago

#8884 - How can I visualize the loss/iteration changes of two different trainings I conducted with YOLOv4 algorithm using input sizes of 256 and 320 in a single graph?

Issue - State: open - Opened by ridvanozdemir 8 months ago

#8883 - gl headers before including cuda_gl_interop.h

Issue - State: open - Opened by SheldonWBM 8 months ago - 1 comment

#8882 - make

Pull Request - State: closed - Opened by Kckarthik2003 8 months ago

#8881 - License for linked resources (i.e. weights file)

Issue - State: closed - Opened by koichiHik 9 months ago - 1 comment

#8880 - Can Darknet train a two-layer LSTM on R8 dataset?

Issue - State: open - Opened by KylePan-01 9 months ago

#8879 - How to set conf threshold for Darknet YOLO AlexeyAB

Issue - State: open - Opened by petrasvobodova2804 9 months ago

#8878 - Questions with mAP chart using YOLOv4

Issue - State: open - Opened by paradise2024 9 months ago - 1 comment

#8877 - I have error when i trying to build Darknet by vs2022

Issue - State: open - Opened by kmliy901054 9 months ago - 1 comment

#8876 - Can't generate libdarknet.so ( library not found for -l:libdarknet.so )

Issue - State: open - Opened by gugo2808 9 months ago

#8875 - increase upscaling

Issue - State: open - Opened by ravi-lourem 10 months ago - 1 comment

#8874 - YOLOV2 ANCHORS

Issue - State: open - Opened by snehashis1997 10 months ago
Labels: Training issue

#8872 - Training Tiny-Yolov3 with a gru layer

Issue - State: open - Opened by Ikhadija-5 11 months ago
Labels: Training issue

#8871 - max_batches in yolov4-custom.cfg

Issue - State: open - Opened by TidelyPom 11 months ago - 2 comments

#8870 - 执行 darknet.exe detector test cfg\coco.data cfg\yolov4.cfg yolov4.weights data\dog.jpg What is the reason why the image is loaded and not marked to the feature target

Issue - State: open - Opened by LuoJianBin22 11 months ago - 1 comment

#8869 - windows搭建环境运行 darknet.exe detector test cfg\coco.data cfg\yolov4.cfg yolov4.weights data\dog.jpg 失败

Issue - State: open - Opened by LuoJianBin22 11 months ago

#8868 - I cannot install on windows

Issue - State: closed - Opened by Waariss 11 months ago - 1 comment

#8867 - Error while training.

Issue - State: open - Opened by yonisoft 11 months ago - 8 comments

#8866 - The confidence level of multiple class models is lower than that of a single model

Issue - State: open - Opened by Nanmei 11 months ago

#8865 - n

Issue - State: closed - Opened by Nanmei 11 months ago

#8864 - @AlexeyAB Is this project no longer maintained?

Issue - State: open - Opened by 1027663760 11 months ago - 2 comments

#8863 - Christian2

Pull Request - State: closed - Opened by christianlanx 11 months ago - 1 comment

#8862 - links in "YOLOv4 model zoo" page are expired? not able to download weights for YOLOv4-Leaky, YOLOv4-Mish?

Issue - State: open - Opened by ridvanozdemir 11 months ago - 2 comments

#8861 - Support CUDA 11.8, Opencv 4.8

Pull Request - State: closed - Opened by c1aude 11 months ago

#8860 - add missing dependencies for openvino

Pull Request - State: closed - Opened by cenit 11 months ago

#8859 - Darknet data load network

Issue - State: open - Opened by Anirudh0603 11 months ago

#8858 - yolov4 (with yolov7.cfg) vs yolov7

Issue - State: open - Opened by jeffersonchua 11 months ago

#8857 - Early fusion of features for multi training

Issue - State: open - Opened by petrasvobodova2804 11 months ago

#8856 - Everything in nan but -show_imgs show correct boxes

Issue - State: open - Opened by karantai 12 months ago

#8855 - OpenCV(4.6.0) /croot/opencv-suite_1691620365762/work/modules/core/src/alloc.cpp:73: error: (-4:Insufficient memory) Failed to allocate 82924982787168 bytes in function 'OutOfMemoryError'

Issue - State: open - Opened by karantai 12 months ago - 1 comment

#8854 - train weight not saved after training 1000 iteration

Issue - State: open - Opened by asyilanrftr 12 months ago - 2 comments
Labels: Training issue

#8853 - batch process images

Issue - State: open - Opened by Miuney 12 months ago

#8852 - How to use Yolo as DLL after building Yolo as C++ DLL-file yolo_cpp_dll.dll

Issue - State: closed - Opened by berkgungor 12 months ago - 6 comments

#8851 - Release version of darknet call crashes

Issue - State: open - Opened by ukoehler 12 months ago

#8850 - mAP too low when training yolov4 tiny custom model

Issue - State: open - Opened by andreluizeng almost 1 year ago - 4 comments
Labels: Training issue

#8849 - How to make yolo v4 output the area and coordinate

Issue - State: open - Opened by db777888999 about 1 year ago

#8846 - 【SOLVED】 Python `darknet.py` ERROR : `Could not find module 'darknet.dll'`, * NO NEED * to downgrade to py 3.7 on Windows 11

Issue - State: open - Opened by 12343954 about 1 year ago

#8845 - Segmentation fault

Issue - State: closed - Opened by TidelyPom about 1 year ago - 2 comments

#8844 - minor code cleaning

Pull Request - State: closed - Opened by cenit about 1 year ago

#8843 - lower disk space usage in CI to remove spurious errors

Pull Request - State: closed - Opened by cenit about 1 year ago

#8842 - cpack update

Pull Request - State: closed - Opened by cenit about 1 year ago

#8841 - YOLO V4 hyperparameter for achieveng higher mAP and accuracy after training

Issue - State: open - Opened by joan2212 about 1 year ago

#8840 - Loss graph lost when training continues on colab

Issue - State: open - Opened by vanquang203546 about 1 year ago - 4 comments

#8839 - I would like to pass both RGB and Greyscale image as part of training data and inference with RGB image, how does the model will perform

Issue - State: open - Opened by gsarathkumar92 about 1 year ago

#8838 - How to get the accuracy of test or valid in classification? Not Top1 or Topk. Such as "- map"

Issue - State: open - Opened by zhen-yan about 1 year ago - 1 comment
Labels: Training issue

#8826 - CUDDA status error

Issue - State: closed - Opened by mizo-1980 about 1 year ago - 2 comments

#8825 - Cuda 11.7 & cudnn8.8.0 unhandled exception

Issue - State: closed - Opened by zirid about 1 year ago - 1 comment

#8821 - convolutional_layer.c:286:24: error: ‘CUDNN_CONVOLUTION_FWD_PREFER_FASTEST’ was not declared in this scope; did you mean ‘CUDNN_CONVOLUTION_FWD_ALGO_FFT’?

Issue - State: closed - Opened by lvshaoqing about 1 year ago - 1 comment

#8820 - if one anchor is asigned to multi label, then should choose the max iou truth label

Pull Request - State: open - Opened by ZouJiu1 about 1 year ago - 5 comments

#8819 - stack smashing detected on Ubantu 20.04

Issue - State: open - Opened by rss1234 about 1 year ago

#8818 - update build script to workaround vcpkg integrated in vs

Pull Request - State: closed - Opened by cenit about 1 year ago

#8817 - Yolov4 and Nvidia GeForce RTX 4090

Issue - State: open - Opened by TidelyPom about 1 year ago

#8816 - Fix typo in kmeansiou.c

Pull Request - State: open - Opened by eltociear about 1 year ago

#8815 - img size wrong

Issue - State: open - Opened by He32100 about 1 year ago - 1 comment

#8814 - Work with data containing "inverted" bounding boxes (from bottom right to top left)

Issue - State: closed - Opened by proplayer2020 about 1 year ago - 2 comments

#8813 - Same cfg & weights , same parameters , give different results . (from 1 detection to 6000)

Issue - State: open - Opened by zirid about 1 year ago - 1 comment

#8812 - Windows darknet.exe use problem

Issue - State: closed - Opened by MarsCaiWORD over 1 year ago

#8811 - YOLOv4-tiny fast prediction, but slow overall progress

Issue - State: open - Opened by JeremyKeusters over 1 year ago - 2 comments

#8810 - Image crop resize is empty

Issue - State: closed - Opened by renatocesargarcia over 1 year ago

#8809 - YOLOv4 output format

Issue - State: open - Opened by aidevmin over 1 year ago - 3 comments

#8808 - modify yolov4 for early stopping

Issue - State: open - Opened by pisenk over 1 year ago

#8807 - How to Show both train and valid map while training

Issue - State: open - Opened by Ataru-Yabe over 1 year ago

#8806 - Community triplet configuration not working with vcpkg

Issue - State: closed - Opened by ColemanKobe over 1 year ago - 1 comment

#8805 - dnn module was not built with cuda backend; switching to cpu

Issue - State: open - Opened by rgund26 over 1 year ago

#8804 - Training on Greyscale images 3 channel vs 1 channel gives different accuracy

Issue - State: open - Opened by WHBSmith over 1 year ago

#8803 - Darknet mrflinxy custom

Pull Request - State: closed - Opened by MrFlinxy over 1 year ago

#8802 - Bug in YOLOv4 darknet_images.py at line 106

Issue - State: open - Opened by robmlee over 1 year ago - 1 comment

GitHub / AlexeyAB/darknet issues and pull requests