Closed bearprin closed 1 year ago
Hi, for train/val sets, we use the same models for pre-training. for the test-time optimization, we use 10 scenes of test set for each category of synthetic room dataset.
room_04: 00000820 00000826 00000827 00000830 00000831 00000832 00000839 00000840 00000988 00000994
room_05: 00000805 00000816 00000838 00000839 00000856 00000869 00000870 00000919 00000990 00000998
room_06: 00000899 00000930 00000931 00000943 00000944 00000945 00000946 00000955 00000960 00000968
room_07: 00000804 00000816 00000831 00000854 00000876 00000917 00000934 00000975 00000989 00000993
room_08: 00000806 00000809 00000810 00000938 00000939 00000940 00000949 00000974 00000981 00000994
Awesome! Thanks for your reply.
Hi, thanks for your work.
I have noticed that you randomly picked 50 shapes for each set (train/val/test) under the synthetic room dataset.
Would you like to share which data you have used (e.g., index)? I want to do the same exp as you.
Looking forward to your reply.