Detectron2: How to inference on images by using PointRend？

[sunnyheart008205]

❓How to inference on images by using PointRend？

Questions like:
When I use this command “python demo/demo_for_pointrend.py --config-file projects/PointRe
nd/configs/InstanceSegmentation/pointrend_rcnn_R_50_FPN_3x_coco.yaml --input datasets/testset/mytest.png --opts MODEL.WEIGHTS ./projects/PointRend/weights/model_final_3c3198_3x_coco.pkl”, the problems following appears.

Traceback (most recent call last):
File "demo/demo_for_pointrend.py", line 75, in
cfg = setup_cfg(args)
File "demo/demo_for_pointrend.py", line 26, in setup_cfg
cfg.merge_from_file(args.config_file)
File "/data/s00378650/InstanceSegmentation/detectron2/detectron2-master/detectron2/config/config.py", line 45, in merge_from_file
self.merge_from_other_cfg(loaded_cfg)
File "/opt/conda/lib/python3.6/site-packages/fvcore/common/config.py", line 121, in merge_from_other_cfg
return super().merge_from_other_cfg(cfg_other)
File "/opt/conda/lib/python3.6/site-packages/yacs/config.py", line 217, in merge_from_other_cfg
_merge_a_into_b(cfg_other, self, self, [])
File "/opt/conda/lib/python3.6/site-packages/yacs/config.py", line 460, in _merge_a_into_b
_merge_a_into_b(v, b[k], root, key_list + [k])
File "/opt/conda/lib/python3.6/site-packages/yacs/config.py", line 460, in _merge_a_into_b
_merge_a_into_b(v, b[k], root, key_list + [k])
File "/opt/conda/lib/python3.6/site-packages/yacs/config.py", line 473, in _merge_a_into_b
raise KeyError("Non-existent config key: {}".format(full_key))
KeyError: 'Non-existent config key: MODEL.ROI_MASK_HEAD.FC_DIM'

Q: How to inference PointRend model by using demo.py?

[ppwwyyxx]

You need to call https://github.com/facebookresearch/detectron2/blob/ee0cbd8c67622ff2753493ce0dcfeb3e3dd9945e/projects/PointRend/train_net.py#L69 to enable it.

[ppwwyyxx]

See also https://colab.research.google.com/drive/1isGPL5h5_cKoPPhVL9XhMokRtHDvmMVL