-
https://github.com/lyuwenyu/RT-DETR/blob/main/rtdetr_pytorch/tools/export_onnx.py Line:78
这个逻辑是不是有点问题,dynamic变成了1的固定batchsize
-
### Search before asking
- [X] I have searched the YOLOv8 [issues](https://github.com/ultralytics/ultralytics/issues) and found no similar feature requests.
### Description
Meta-class to lo…
-
### 问题确认 Search before asking
- [X] 我已经搜索过问题,但是没有找到解答。I have searched the question and found no related answer.
### 请提出你的问题 Please ask your question
Hi authors,
I am interested in RT-DETR, and I …
-
您好,看到在类似的detr方法中有使用swin替换之前的resnet主干网络,在coco上效果可以提升很多,我们有做过类似的实验吗,在降低耗时优先级的情况下是否有效果的提升
-
作者您好,对于`rtdetr_r18vd_6x_coco`,如果我将`num_queries`由300改成100,训练时其他参数是否需要跟着修改,比如`num_denoising`是否需要相应的缩小3倍,100—>33,还有其他参数需要调整吗?
-
已star,您好,请问下:源码pytorch版本,按照您的步骤设置好了。train如何提高bath_size和num_workers,显存和gpu利用率都不变,并且发现cpu在训练,爆内存是设置太高了的问题。似乎没用gpu,我改如何处理这个问题呢?
-
我按照requirements.py 安装后出现以下报错 搜索看到可能与torch的版本有关。
Start training
Load PResNet50 state_dict
Initial lr: [1e-05, 0.0001, 0.0001, 0.0001]
loading annotations into memory...
Done (t=11.22s)
creating …
-
### Search before asking
- [X] I have searched the Ultralytics YOLO [issues](https://github.com/ultralytics/ultralytics/issues) and [discussions](https://github.com/ultralytics/ultralytics/discussion…
733ly updated
3 weeks ago
-
### 问题确认 Search before asking
- [X] 我已经搜索过问题,但是没有找到解答。I have searched the question and found no related answer.
### 请提出你的问题 Please ask your question
我发现PaddleYOLO里的算法,只要用了MosaicPerspective数据增强(如yo…
-
转换的命令行如下:
trtexec --onnx=rtdetr_hgnetv2_l_6x_coco_quant_dynamic.onnx --saveEngine=rtdetr_hgnetv2_l_6x_coco_quant_dynamic.engine --minShapes=images:1x3x640x640 --optShapes=images:16x3x640x640 --maxSha…
zhDai updated
4 months ago