github
/
darknet
mirror of https://github.com/AlexeyAB/darknet.git
1
0
Fork 0
Code Issues Releases Wiki Activity
1758 Commits
1 Branch
7 Tags
380 MiB
Tag: Branch: Tree: 3d9aa2af47
Branches Tags
${ item.name }
Create tag ${ searchTerm }
Create branch ${ searchTerm }
from '3d9aa2af47'
${ noResults }
Commit Graph

219 Commits (3d9aa2af4718a3bd5bbe23de2022987cb767e9c5)

Author SHA1 Message Date
AlexeyAB 3d9aa2af47 `./darknet detector calc_anchors` calculates both anchors= and counters_per_class=
5 years ago
AlexeyAB 4cd21785fc Minor fix
5 years ago
AlexeyAB 9919612006 Added attention during training: [net] adversarial_lr=0.05 attention=1
5 years ago
AlexeyAB 978c2e00ed self-adversarial training
5 years ago
AlexeyAB 39d79785ba Minor fix
5 years ago
AlexeyAB b3c9f373d1 adversarial attack - minor improvement
5 years ago
AlexeyAB aa4f9975a7 adversarial attack - fixed
5 years ago
AlexeyAB 5e904d5b1e adversarial attack - experimental
5 years ago
AlexeyAB 864d1062f8 Minor fix
5 years ago
Muhammad Maaz 8800212faa Adds support to print remaining training time on console if darknet is not compiled with OpenCV
5 years ago
Muhammad Maaz bd7690b019 Fix: Correct remaining training time in case someone resume training (src/detector.c)
5 years ago
AlexeyAB 5c2ddd301e Added param to cfg: [net] gaussian_noise=40
5 years ago
Muhammad Maaz 1a6dd593c5 Fix: Use iteration instead of i for current iteration.
5 years ago
Muhammad Maaz b77add049e Logic Change: Implements rolling avg for remaining time calculation in detector.c
5 years ago
AlexeyAB 92e6e8eece Minor fix
5 years ago
AlexeyAB 0c8e9ce4b2 Minor fix
5 years ago
AlexeyAB d11caf486d Fixed BiFPN and label_smothing for Detection
5 years ago
AlexeyAB 5e54523960 Added loss_scale for Mixed-Precision on Tensor-Cores
5 years ago
AlexeyAB 1a56ef588a Added flag -cuda_debug_sync for checking CUDA errors
5 years ago
AlexeyAB a567a806a6 Set cur_iteration = 0; initially
5 years ago
Muhammad Maaz 1174bb7417 Add support to ignore first 50 iterations in calculating the remaining training time.
5 years ago
Muhammad Maaz d124d552c0 Add feature to show approx. remaining training time to training chart.
5 years ago
Muhammad Maaz 324cc881ee Add support to display time remaining on the training chart.
5 years ago
AlexeyAB c814d56ec1 Added [net] dynamic_minibatch=1 for increasing mini_batch_size when random=1 is used
5 years ago
AlexeyAB 8d87df2388 Minor fix
5 years ago
AlexeyAB 8dc4833e7f Minor fix
5 years ago
Ksenya Ryabinova f441b3025a Added -chart argument
5 years ago
AlexeyAB ecf1aeb7e7 -benchmark_layers for backward too (for training)
5 years ago
AlexeyAB 04f78ddf72 Fixed bugs from PR
5 years ago
AlexeyAB 2247c5fcd7 minor fixes
5 years ago
AlexeyAB a66bc8d6ff Resizing Detector: random=1 (/1.4 - x1.4), random=3 (/3.0 - x3.0), random=0 (no resizing). Added param: resize_step=32 to the [net] section: use resize_step=128 for 7 subsampling layers.
5 years ago
cyy 9d2e6f9a23 fix compiling errors
5 years ago
AlexeyAB 2652263727 Added flags -benchmark and -benchmark_layers for: ./darknet detector demo
5 years ago
AlexeyAB 2c08a20b22 Added flag -benchmark_layers for: classifier demo, detector test, detector train
5 years ago
AlexeyAB 81c5bcea92 Avg-loss will be saved to the 2 images: chart.png & chart_your_cfg_name.png
5 years ago
AlexeyAB 1626f9d495 Some fixes
6 years ago
AlexeyAB 318919e1cb Added label_smooth_eps=0.1 for [yolo] and [Gaussian_yolo] layer for Label Smoothing
6 years ago
AlexeyAB aaba21bdd3 Added ./darknet detector valid for KITTI and BDD (were previously MSCOCO and ImageNet)
6 years ago
AlexeyAB fbdac785d2 Added HTTP POST sender with detected objects, use: ./darknet detector demo ... -http_post_host webhook.site/898bbd9b-0ddd-49cf-b81d-1f56be98d870
6 years ago
AlexeyAB 8a3a7ccac6 Added `./darknet detector demo ... -time_limit_sec 10` to limit detection time
6 years ago
AlexeyAB f1ffb09d8b Fixed minor bugs and memory leaks
6 years ago
AlexeyAB b4fcb8b294 Fixed: revert disabled mAP calculation during training.
6 years ago
AlexeyAB 0d98f20be1 You can train with mini_batch=128 (batch=256 subdivisions=2) on GPU 8 GB VRAM + 128 GB CPU-RAM and more. Set in cfg-file: optimized_memory=3 and workspace_size_limit_MB=2000 or 4000
6 years ago
AlexeyAB 4f70fc14e9 Added DIoU and CIoU to [yolo] and [Gaussian_yolo]
6 years ago
AlexeyAB 77e60aecfe Minor fix
6 years ago
AlexeyAB 3555beb914 Fixed some memory leaks in secondary functions
6 years ago
AlexeyAB e6486ab594 Use ignore_thresh only if class_id matched. Temporary changed Assisted_Excitation (reduces background activations rather than enhancing objects activations). Added antialiasiong=2 for 2x2.
6 years ago
cyy 41144ba358 fix calloc realloc failure
6 years ago
cyy 011eed632d fix calloc realloc failure
6 years ago
cyy 81ab7adc28 fix warnings
6 years ago