Update Dockerfile

[ErikValle]

You can test it by running the following commands:

docker build -t vitpose:latest .
docker run -it --name vitpose_test --gpus all vitpose:latest

Inside your container:

python demo/topdown_demo_with_mmdet.py demo/mmdetection_cfg/faster_rcnn_r50_fpn_coco.py https://download.openmmlab.com/mmdetection/v2.0/faster_rcnn/faster_rcnn_r50_fpn_1x_coco/faster_rcnn_r50_fpn_1x_coco_20200130-047c8118.pth configs/body_2d_keypoint/topdown_heatmap/coco/td-hm_hrnet-w48_8xb32-210e_coco-256x192.py https://download.openmmlab.com/mmpose/top_down/hrnet/hrnet_w48_coco_256x192-b9e0b3ab_20200708.pth --input demo/resources/demo.mp4 --output-root vis_results

[LukaMarkicevic]

Did you manage to get ViTPose working as well?
You are using HRNet in the command.

[bpezet]

Did you manage to get ViTPose working as well? You are using HRNet in the command.

Any updates on using ViTPose models ?

[ErikValle2]

I will try it later, keep you posted.

[FETHl]

Any updates on using the ViTPose or ViTPose+ models, please? I need them, but I haven’t been able to install ViTPose correctly yet. HRNet is working fine, but ViTPose is not set up so far.

[LukaMarkicevic]

I solved it by doing what jun297 says in #97.
The main point is that you need torch 1.9.0 with cuda 11.1 and a specific version of mmcv-full and not mmcv.
Maybe it can help someone.