07ab35222c
* add v8 training head * add training config * update * update1 * update all config * update all config * update metafile * add head loss * update readme * iou2ciou in tal, add config * change init layer * update * update * add clip_grad * update max_aspect_ratio * add init * fix aug config * update forward * add official v8 loss * add official v8 loss * 同步预测方式 * fix loss bug * delete useless code * 替换iou，但还有问题 * fix bug * fix empty * add UT * add UT * update readme * replace iou_cal * add use_ciou docstr * update metafile * update metafile * update metafile Co-authored-by: huanghaian <huanghaian@sensetime.com> Co-authored-by: hha <1286304229@qq.com> |
||
|---|---|---|
| .. | ||
| README.md | ||
| metafile.yml | ||
| yolov8_l_syncbn_fast_8xb16-500e_coco.py | ||
| yolov8_m_syncbn_fast_8xb16-500e_coco.py | ||
| yolov8_n_syncbn_fast_8xb16-500e_coco.py | ||
| yolov8_s_syncbn_fast_8xb16-500e_coco.py | ||
| yolov8_x_syncbn_fast_8xb16-500e_coco.py | ||
README.md
YOLOv8
Abstract
Ultralytics YOLOv8, developed by Ultralytics, is a cutting-edge, state-of-the-art (SOTA) model that builds upon the success of previous YOLO versions and introduces new features and improvements to further boost performance and flexibility. YOLOv8 is designed to be fast, accurate, and easy to use, making it an excellent choice for a wide range of object detection, image segmentation and image classification tasks.
performance
YOLOv8-P5 model structure
Results and models
COCO
| Backbone | Arch | size | SyncBN | AMP | Mem (GB) | box AP | Config | Download |
|---|---|---|---|---|---|---|---|---|
| YOLOv8-n | P5 | 640 | Yes | Yes | 2.8 | 37.2 | config | model | log |
| YOLOv8-s | P5 | 640 | Yes | Yes | 4.0 | 44.2 | config | model | log |
| YOLOv8-m | P5 | 640 | Yes | Yes | 7.2 | 49.8 | config | model | log |
Note
In the official YOLOv8 code, the bbox annotation, random_perspective and copy_paste data augmentation in COCO object detection task training uses mask annotation information, which leads to higher performance. Object detection should not use mask annotation, so only box annotation information is used in MMYOLO. We trained the official YOLOv8s code with 8xb16 configuration and its best performance is also 44.2. We will support mask annotations in object detection tasks in the next version.
- We use 8x A100 for training, and the single-GPU batch size is 16. This is different from the official code, but has no effect on performance.
- The performance is unstable and may fluctuate by about 0.3 mAP and the highest performance weight in
COCOtraining inYOLOv8may not be the last epoch. The performance shown above is the best model. - We provide scripts to convert official weights to MMYOLO.
SyncBNmeans use SyncBN,AMPindicates training with mixed precision.