ValueError: too many values to unpack (expected 2)

[MacBook-Pro-gala]

python train_ft.py --config ../dev_scripts/spidr/lego.ini --run_mode=sdf

xyz_world_all torch.Size([91388555, 3]) torch.Size([91388555, 1]) torch.Size([91388555])
%%%%%%%%%%%%% getattr(dataset, spacemin, None) None
vishull_mask torch.Size([91388555])
alpha masking xyz_world_all torch.Size([68219605, 3]) torch.Size([68219605, 1])
xyz_residual torch.Size([68219605])
min_idx torch.Size([1079024])
after voxelize: torch.Size([1079024, 3]) torch.Size([1079024, 1])
0%| | 0/549 [00:00<?, ?it/s]Traceback (most recent call last):
File "train_ft.py", line 679, in
main()
File "train_ft.py", line 146, in main
img_lst, c2ws_lst, w2cs_lst, intrinsics_all, HDWD_lst = gen_points_filter_embeddings(train_dataset, visualizer, opt)
File "/code/oldSPIDR/SPIDR/run/ft_helper.py", line 111, in gen_points_filter_embeddings
model.query_embedding(HDWD_lst[i], torch.as_tensor(cam_xyz_all[i][None, ...], device="cuda", dtype=torch.float32),
File "/code/oldSPIDR/SPIDR/run/../models/mvs_points_volumetric_model.py", line 321, in query_embedding
return self.net_mvs.query_embedding(HDWD, cam_xyz, photometric_confidence, img_feats, c2ws, w2cs, intrinsics, cam_vid, pointdir_w=pointdir_w)
File "/code/oldSPIDR/SPIDR/run/../models/mvs/mvs_points_model.py", line 240, in query_embedding
, view_ids = feat_str.split("")
ValueError: too many values to unpack (expected 2)
end loading

[liuyuwan233]

Hi！
I had the same problem as you. I solved this problem by rewrting this setting "appr_feature_str0 = [imgfeat_0_0123, dir_0, point_conf]" in the ../dev_scripts/w_n360/lego.ini.
I hope this helps you.