harrymore

cannot use onnxruntime-gpu to infer model

10

when I run detection model, only cpu is used. my package version： insightface==0.5 onnx==1.10.2 onnxruntime-gpu==1.9.0 even if I uninstalled insightface and installed it again.

关于backbone

2

你好，请问在论文中提到的backbone，有vgg16 tiny，这个网络应该不是完整的vgg16吧，你们是提取哪一层作为fetrue map呢，是stride=1/8那层的输出吗？

目录预览.md里的第十五章是正则化，但是实际是“异构计算， GPU和框架选型指南”

修改了gpushare-device-plugin-ds 中的 --memory-unit=MiB，重启之后，使用`kubectl inspect gpushare`命令显示如下： > NAME IPADDRESS GPU0(Allocated/Total) GPU Memory(MiB) > k8s-node0 192.168.1.3 0/6144 0/6144 > k8s-node1 192.168.1.4 0/8192 0/8192 > k8s-node2 192.168.1.5 0/6144 0/6144 > ---------------------------------------------------------- > Allocated/Total...

[BUG] Node (Reshape_2173) Op (Reshape) [ShapeInferenceError] Dimension could not be inferred: incompatible shapes

1

**Describe the bug** I use the mmdeploy framework to convert the Cascade Mask R-CNN (FPN) model to an ONNX model. After optimization with onnxsim, I perform inference using onnxruntime,then the...