mmpose on a virtual server

[IssaIssa1]

Hello,
I am trying to run mmpose predict on a virtual server using jupyter notebook. I followed the colab tutorial I had to set up proxy as follows for the instalations

import os os.environ['https_proxy'] = "http:https://proxy:80" os.environ['http_proxy'] = "http:https://proxy:80"

But when I run the prediction it tries to download the model as follows:

Downloading: "https://download.openmmlab.com/mmpose/top_down/hrformer/hrformer_base_coco_384x288-ecf0758d_20220316.pth" to /home/hpc/iwso/iwso067h/.cache/torch/hub/checkpoints/hrformer_base_coco_384x288-ecf0758d_20220316.pth

But it returns a time out error

URLError: <urlopen error [Errno 110] Connection timed out>

I suspect it's proxy issue but I couldn't fix it. I'd appreciate your help

[ly015]

MMPose will automatically download the checkpoint to a temporal folder at the first time it is used in inference. Are you face any problems at this step?

[IssaIssa1]

I updated my issue. I added the error message. I am getting time out error exactly on the step where the model is automatically downloading.

[ly015]

It looks like the download link can not be accessed from your virtual server. Is it possible to manually upload the checkpoint to your server?

[IssaIssa1]

I have access to the server but I am experimenting with different models so it would be faster and more efficient if it can be fixed through code. I was trying to add proxy to mmpose source code but no luck fixing it yet.