Open BowieHsu opened 1 year ago
BowieHsu
commented
1 year ago 大家好，同步一下最新的进展：我们已经上传了yolov5l/yolov5m/yolov5s/yolov7的coco预训练模型，放置在1.0的release文件中了，可以直接用detect和val代码进行渲染和计算准确率，欢迎试用和提出相应问题；
Hello everyone, here's an update: We have uploaded pre-trained YOLOv5l/YOLOv5m/YOLOv5s/YOLOv7 models on COCO dataset. You can find these models in the 1.0 release files and directly use detect and val codes to render and compute accuracy. Feel free to try and ask any relevant questions.
BowieHsu
commented
1 year ago 最新进展：我们上传了是用SimOTA来训练YOLOv7s的脚本和coco模型，方便各位用户通过比较configs/sup/public/yolov7.cfg和configs/sup/public/yolov7s-simota.cfg来了解如何将一个anchor based的检测器变成anchor free的检测器，这里面只会修改一点点配置文件，欢迎大家试用。
Latest update: We have uploaded the training script and coco pretrain model for YOLOv7s with SimOTA label assigner. Users can now compare configs/sup/public/yolov7.yaml and configs/sup/public/yolov7s-simota.yaml to learn how to convert an anchor-based detector to an anchor-free detector with just a few modifications to the configuration file. We encourage everyone to try it out.
请允许我介绍一下3月的更新计划——1.我们将会在验证后上传在readme中提到的这些coco预训练模型，这样可以进一步方便大家调试和使用我们的算法库，也能更快地获得较好的半监督训练效果；2.我们的半监督方案还无法适应任意的数据分布，所以我们希望收集大家在实际使用中遇到的各种情况，请随时提出issue来共同改进；
Let me introduce the update plan for March: 1. After verification, we will upload the coco pre-trained models mentioned in readme, which will further facilitate everyone to debug and use our algorithm library, and can also get better semi-supervised training effect faster; 2. Our semi-supervised scheme can not adapt to any data distribution, so we hope to collect various situations encountered by everyone in actual use, please feel free to submit issues to improve together;