I run evc -t test -c configs/projects/realtime4dv/rendering/4k4d_sport1.yaml,configs/specs/eval.yaml in NHR dataset and find that the code test 0th, 25th, and 50th input views (200 frames). Is this consistent with the experiment in the paper? In addtion, the code(basketball) test 0th, 26th, and 52th input views(194 frames).
I run
evc -t test -c configs/projects/realtime4dv/rendering/4k4d_sport1.yaml,configs/specs/eval.yaml
in NHR dataset and find that the code test 0th, 25th, and 50th input views (200 frames). Is this consistent with the experiment in the paper? In addtion, the code(basketball) test 0th, 26th, and 52th input views(194 frames).