harrymore

可以尝试把防火墙关掉试试

This plugin does not do memory isolation,so you should only do this manually,in your code or environment variables

明白，谢谢。另外，问一个问题，请问你们的GPP，每个组合并完后是多少通道呢？

尝试使用亲和力去实现

> the version of onnxruntime-gpu library should match your cuda version. U R right, the version of my cuda is 10.1,I degrade onnxrutime-gpu and insightface to suitable version and did...

> insightface can support CPU inference? and give a code example ? tks. and it's not support python 3.8.8. casuse the version of my cuda doesn't match onnxruntime-gpu, so when...

@ninenox-dev please install suitable version of onnxruntime according to the versions of cuda and cudnn,you can find the table by the link: https://onnxruntime.ai/docs/execution-providers/CUDA-ExecutionProvider.html

我当时爬到的数据也是不全的，只爬到豆瓣放出来的有限数据，很多也爬不到

my case is that if I set MiB，use commad"kubectl inspect gpushare" display GPU with MiB unit,but when I apply for gpu in pod, it remand me: 0/3 nodes are available:...