YOLOv5-Multibackbone-Compression icon indicating copy to clipboard operation
YOLOv5-Multibackbone-Compression copied to clipboard

YOLOv5 Series Multi-backbone(TPH-YOLOv5, Ghostnet, ShuffleNetv2, Mobilenetv3Small, EfficientNetLite, PP-LCNet, SwinTransformer YOLO), Module(CBAM, DCN), Pruning (EagleEye, Network Slimming), Quantizat...

Results 58 YOLOv5-Multibackbone-Compression issues
Sort by recently updated
recently updated
newest added

@Gumpest 应该检测的是加完cbam注意力的那一层吧。

大佬您好,请问hyp.visdrone.yaml里面的超参数是用yolov5的超参数进化来设置的吗?能否适用于其它网络模型?谢谢

EagleEye搜索最优子网不是很清楚,可以分享一下经验吗?我迁移到其他模块不能运行,比如MobileNetv2,MobileNetv3等模型,希望你能指导,非常感谢。

您好,请问您base model输入图片大小和Higher Precision Model输入图片大小不一样,没有用到控制变量法,这在我们后续写论文中可以这样做嘛

Hello author, if you run standard models such as yolov5m with your code, will the effect be exactly the same as running yolov5m with the official code? Or does your...

参数设置,官方预训练模型,batchsize=2,图片大小608 监控了显卡显存是一直积累的。但是使用yolox等训练是没问题的。 大佬,能否看一下报错 autoanchor: Analyzing anchors... anchors/target = 5.85, Best Possible Recall (BPR) = 0.9984 Image sizes 608 train, 608 val Using 2 dataloader workers Logging results to runs/train/exp8 Starting...

Thanks a lot for this awesome repos. It seems that export.py is the same as ultralytics's yolov5 repo, which is suitable for standard yolov5 models. However, when exporting TPH model...

您好,大佬,我现在非常困惑该如何进行调参,有没有调参的经验或者指南可以借鉴下? 我看一篇论文训练DIOR数据集9个类别,能达到0.807的mAP,而我使用默认的hyps.scratch.yaml才只达到了0.745的mAP. ![image](https://user-images.githubusercontent.com/46252413/144588020-c48e5fc6-9c08-46ab-83f4-ae0732e9386f.png) ![image](https://user-images.githubusercontent.com/46252413/144588045-ab423d5d-2329-4dbd-bc01-67e7f3521be3.png) 对于 桥 这种我觉得很长很大的目标反而最终的mAP不太高。 ![image](https://user-images.githubusercontent.com/46252413/144591479-2e8e7faa-a72e-469a-b8d2-7be29da19a81.png)

您好,请问为什么我使用yolov5x-TPH训练小目标数据集的最终map只有95.5,而使用yolov5x训练同样的数据集最终map有99,请问这是什么原因呢?