Open LiuShuai086 opened 10 months ago
same problem, i run this, from offical scripts: ./scripts/runall_syn_sdf.sh
CUDA_VISIBLE_DEVICES=7 python main.py data/nerf_synthetic/lego/ --workspace trial_syn_sdf_lego/ -O --bound 1 --scale 0.8 --dt_gamma 0 --stage 0 --sdf
I find the quantitative results of the sdf mode are generally worse than the non-sdf mode. Would you or others provide some advice about the configuration of the sdf mode? Sincerely waiting for your kind help.