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

大佬您好,关于BIFPN的实现在哪个文件里面呢,如果我要将其用在YOLOv5s模型里面,应该修改一些什么参数呢

return flops * img_size[0] / stride * img_size[1] / stride UnboundLocalError: local variable 'stride' referenced before assignment

可以分享最新的YOLOv5l-GhostNet的预训练权重吗?我用网盘上面的预训练文件训练不了,好像是文件损坏了,大佬可以再分享一下吗?

我训练了coco128数据集,发现bnweights好像没有效 ![2022-07-22 13-37-34屏幕截图](https://user-images.githubusercontent.com/34853551/180370488-5b2bd343-f10d-48da-a278-26c873911e41.png) 果

使用Ghost代替YOLOV5的主干网络之后,我看到作者的实验结果是map降低了,那请问FPS有提高吗

如题,大佬,没有用大佬的正常训练的,是拿U神的原版v6版本训练出来的权重文件,拿来直接搜索最优子网然后微调不知道有没有影响,因为看到大佬的剪枝yaml里面有许多1.0的参数

![image](https://user-images.githubusercontent.com/84652756/174983148-397e1187-9c82-4d8d-b656-60eedbe718db.png) 请问怎么解决?Traceback (most recent call last): File "pruneEagleEye.py", line 127, in rand_prune_and_eval(model,ignore_idx,opt) File "pruneEagleEye.py", line 59, in rand_prune_and_eval pruned_yaml = update_yaml(pruned_yaml, model, ignore_conv_idx, maskconvdict, opt) File "/hy-nas/yolov5-multibackbone/utils/prune_utils.py", line 263, in...

from models.SwinTransformer import SwinTransformerLayer出错, Traceback (most recent call last): File "train.py", line 24, in import test # import test.py to get mAP after each epoch File "/content/drive/My Drive/yolov5-5.0/test.py", line 12,...

@Gumpest 大佬,在使用MQBench,加载网络就出错了,能帮忙看看原因吗? 加载到第3层出现如下错误 index 0 is out of bounds for dimension 0 with size 0 关闭量化参数quantize,就正常,请教