AlexeyAB/darknet issues and pull requests

#8364 - optical flow tracking， Makefile:169: recipe for target 'libdarknet.so' failed

Issue - State: closed - Opened by Dennyyang7 over 2 years ago

#8362 - Error: cuDNN isn't found FWD algo for convolution. - Please help!

Issue - State: closed - Opened by JSarko27 over 2 years ago - 3 comments

#8343 - mAP 0 during training, but it works when executing map command alone not meanwhile with training

Issue - State: open - Opened by YCAyca over 2 years ago - 2 comments

#8319 - Update yolov4-custom.cfg

Pull Request - State: open - Opened by Basuki74 over 2 years ago

#8302 - Could you check parameters in cudnnSetPooling2dDescriptor() of maxpool_layer?

Issue - State: open - Opened by Namdosa almost 3 years ago - 2 comments

#8297 - Model gives zero error and zero accuracy after training

Issue - State: open - Opened by ankit41093 almost 3 years ago - 1 comment
Labels: Training issue

#8281 - What does yolov4.conv.81 mean?

Issue - State: closed - Opened by Turkishvein almost 3 years ago - 2 comments

#8197 - changes to crop detections in video

Pull Request - State: open - Opened by rg321 almost 3 years ago

#8178 - Added script to generate text file with dataset images relative paths.

Pull Request - State: open - Opened by KyryloAntoshyn almost 3 years ago - 5 comments

#8172 - 5 second delay with rtsp stream

Issue - State: closed - Opened by maddios almost 3 years ago - 7 comments
Labels: Solved

#8160 - Improving model performance in vary sizes of the same object

Issue - State: open - Opened by InbarKedem almost 3 years ago - 15 comments

#8110 - fix-for-extreme-verbose-while-training

Pull Request - State: open - Opened by nagi1995 about 3 years ago

#8081 - small objects yolov4-csp

Issue - State: open - Opened by tcdnm about 3 years ago - 4 comments

#8048 - Added custom object detection folder with all required files

Pull Request - State: open - Opened by shanky1947 about 3 years ago

#8047 - Added video detection code

Pull Request - State: open - Opened by mihir-agarwal0211 about 3 years ago

#7955 - Issue run image classification

Issue - State: open - Opened by SokPhanith about 3 years ago - 5 comments

#7947 - train well map zero with p5 and csp swish

Issue - State: open - Opened by mnoroozii about 3 years ago - 8 comments

#7823 - No prediction result with GPU version

Issue - State: closed - Opened by binarybehemoth over 3 years ago - 2 comments
Labels: Solved

#7644 - deprecate msbuild solutions

Pull Request - State: open - Opened by cenit over 3 years ago - 16 comments

#7638 - Random -nan loss after evaluation

Issue - State: open - Opened by ekesdf over 3 years ago - 3 comments
Labels: Training issue

#7608 - About ciou and diou loss.

Issue - State: open - Opened by SpongeBab over 3 years ago - 7 comments

#7599 - Darknet installation on Windows10 with RTX 3090 - CUDA 11.2 and cudNN 8

Issue - State: closed - Opened by MiguelEsteve over 3 years ago - 8 comments

#7596 - Novel EIoU Loss

Issue - State: open - Opened by Dimfred over 3 years ago - 2 comments
Labels: Feature-request

#7282 - Updated the code that allows to save the cropped image in result_img directory

Pull Request - State: open - Opened by shubham-shahh over 3 years ago - 19 comments

#7252 - Darknet map calculation is extremely slow on RTX3090

Issue - State: open - Opened by mayur-ag over 3 years ago - 16 comments

#7153 - cuDNN Error while training -map

Issue - State: closed - Opened by niemiaszek almost 4 years ago - 21 comments

#7126 - Allowed training with data tagged on more classes than configuration.…

Pull Request - State: open - Opened by gnassi almost 4 years ago

#7100 - remove redundant memory traffic

Pull Request - State: open - Opened by FindHao almost 4 years ago - 2 comments

#6836 - CUDNN_STATUS_BAD_PARAM

Issue - State: open - Opened by lostagex almost 4 years ago - 13 comments

#6811 - Add rotation support in detector train mode

Pull Request - State: open - Opened by robosina almost 4 years ago - 2 comments

#6760 - Free image memory in Python darknet.py

Pull Request - State: open - Opened by betegon about 4 years ago

#6735 - Add RTMP stream function with FFMPEG

Pull Request - State: open - Opened by edwardxliu about 4 years ago - 9 comments

#6717 - Fix darknet_video.py to so it's not off by 1 frame

Pull Request - State: open - Opened by mazatov about 4 years ago

#6666 - Fix image.c resize_min return origin image may cause double free

Pull Request - State: open - Opened by tenpend about 4 years ago - 1 comment

#6535 - CUDA Error: invalid device symbol: File exists

Issue - State: open - Opened by joe-tsai-wj about 4 years ago - 1 comment

#6528 - Different Results on GPU vs CPU

Issue - State: open - Opened by VisionEp1 about 4 years ago - 4 comments
Labels: Likely bug

#6426 - How to detect smaller objects with custom object datasets

Issue - State: closed - Opened by praneeth0609 about 4 years ago - 56 comments

#6407 - Add realsense

Pull Request - State: open - Opened by juhong-rdv about 4 years ago - 1 comment

#6258 - near life stream view independent from input frame rate

Issue - State: open - Opened by UweW about 4 years ago - 2 comments
Labels: Feature-request

#6121 - Support big-endian systems

Pull Request - State: closed - Opened by rsaxvc over 4 years ago

#5941 - Update stb libraries to latest from upstream

Pull Request - State: open - Opened by imaami over 4 years ago - 1 comment
Labels: ToDo

#5920 - Repo Claims To Be YOLOv5

Issue - State: closed - Opened by danielbarry over 4 years ago - 93 comments

#5879 - small fix for cost layer

Pull Request - State: open - Opened by LEA0317 over 4 years ago

#5837 - OpenCV memory error while trying to train with efficientnet-lite3.cfg

Issue - State: open - Opened by Suraj520 over 4 years ago - 1 comment

#5742 - Draft of test time augmentation.

Pull Request - State: open - Opened by LucasSloan over 4 years ago

#5654 - added OO python detector and venv setup file.

Pull Request - State: open - Opened by sarafalamaki over 4 years ago - 2 comments

#5650 - train with efficientnet-lite3.cfg

Issue - State: open - Opened by leijuzi over 4 years ago - 1 comment

#5624 - Update image.c

Pull Request - State: closed - Opened by SaiquaIslam over 4 years ago - 1 comment

#5620 - Amend best_map during training to emulate mAP@[.5:.95]

Pull Request - State: closed - Opened by willbattel over 4 years ago - 3 comments

#5599 - Save only the detected image in darknet video

Issue - State: closed - Opened by ismael694 over 4 years ago - 2 comments

#5460 - Detect mat without opencv

Pull Request - State: open - Opened by SpiRaiL over 4 years ago

#5415 - For testing and saving videos using darknet_video.py

Pull Request - State: open - Opened by Sambhav300899 over 4 years ago

#5181 - Add ability to pause video by pressing 'space'

Pull Request - State: open - Opened by dianaRusak over 4 years ago

#5079 - Comparison of some models on CPU vs VPU (neurochip) vs GPU

Issue - State: open - Opened by AlexeyAB over 4 years ago - 28 comments
Labels: Explanations

#4892 - About the license

Issue - State: closed - Opened by lyqqhgzj over 4 years ago - 11 comments

#4560 - JNA support for darknet

Pull Request - State: open - Opened by peterdn1 almost 5 years ago - 4 comments

#4540 - added counters_per_class Documentation

Pull Request - State: open - Opened by HagegeR almost 5 years ago

#4469 - Label of images without any object

Issue - State: closed - Opened by Adaverse almost 5 years ago - 3 comments

#4459 - Action Recognition

Issue - State: open - Opened by jamessmith90 almost 5 years ago - 11 comments
Labels: question

#4447 - EfficientNetb0-Yolo speed slow

Issue - State: open - Opened by Code-Fight almost 5 years ago - 19 comments
Labels: question

#4162 - Segmentation fault while runnnig yolo on multiple images and try to save it in text file.

Issue - State: closed - Opened by mechamrit almost 5 years ago - 3 comments

#3893 - make directories of multiple layers

Pull Request - State: closed - Opened by keineahnung2345 about 5 years ago

#3856 - How to improve false positive of custom objects?

Issue - State: closed - Opened by lifeel about 5 years ago - 1 comment

#3723 - Evo: add output to a json file option when running on a video

Pull Request - State: closed - Opened by tdurand about 5 years ago - 3 comments

#3692 - Cleanup-remove an unused variable in gen_anchors.py

Pull Request - State: open - Opened by developer0hye about 5 years ago - 5 comments

#3465 - Best weight check between training sessions and exported set_batch_network

Pull Request - State: open - Opened by keko950 over 5 years ago - 14 comments

#3343 - Linux Packages

Issue - State: closed - Opened by acxz over 5 years ago - 13 comments
Labels: ToDo

#3313 - Weights File not saving

Issue - State: open - Opened by mad0511 over 5 years ago - 22 comments

#3309 - Added new Python utilities

Pull Request - State: open - Opened by angeligareta over 5 years ago - 2 comments

#3082 - Found a problem

Issue - State: closed - Opened by ANewAnonymous over 5 years ago - 2 comments

#2830 - No static detector

Pull Request - State: open - Opened by randellhodges over 5 years ago - 3 comments

#2468 - How to train to detect lisence plate ?

Issue - State: closed - Opened by masoudnick over 5 years ago - 17 comments

#2024 - cannot compile any of the .sln files

Issue - State: closed - Opened by xnuken almost 6 years ago - 1 comment

#1939 - Created headless detector test phase with JSON predictions and bounding boxes for every input image; all outputted to single JSON file

Pull Request - State: open - Opened by gpsmit almost 6 years ago - 7 comments

#1885 - added writing progress to file in backup directory

Pull Request - State: open - Opened by haimha6 almost 6 years ago

#911 - Average loss value not going below 1

Issue - State: closed - Opened by r3s over 6 years ago - 5 comments

#478 - Bug in blas.c smooth_l1_cpu

Issue - State: closed - Opened by wilstro over 6 years ago - 2 comments
Labels: wontfix

GitHub / AlexeyAB/darknet issues and pull requests