save ply "parameters" object is not callable

[saliteta]

When running train_affinity_features.py, it will raise error at line 227 when saving ply:
see train_affinity_features.py line 227

The save_feature will trigger save_ply

Where get_point_features is a parameters.

Here is the error and log

(laga) (base) bxiong@anim-deep3-ub:~/workspace/LaGa$ python train_affinity_features.py -m output/3c591536-2 
Looking for config file in output/3c591536-2/cfg_args
Config file found: output/3c591536-2/cfg_args
Optimizing output/3c591536-2
Loading trained model at iteration 30000, None [27/06 10:07:55]
Allow Camera Principle Point Shift: False [27/06 10:07:55]
Reading camera 299/299 [27/06 10:08:03]
Loading Training Cameras [27/06 10:08:03]
Loading Test Cameras [27/06 10:08:06]
Number of 3D Gaussians: 728650 [27/06 10:08:07]
Gaussian adjustable: False [27/06 10:08:07]
Training progress:  23%|███████████████▋                                                   | 7000/30000 [19:39<1:04:54,  5.91it/s, RFN=0.854, Pos cos=0.930, Neg cos=-0.029, Multi Lvl Loss=-2.228, Loss=-2.207]Traceback (most recent call last):
  File "/home/bxiong/workspace/LaGa/train_affinity_features.py", line 264, in <module>
    training(lp.extract(args), op.extract(args), pp.extract(args), args.iteration, args.downsample)
  File "/home/bxiong/workspace/LaGa/train_affinity_features.py", line 228, in training
    scene.save_feature(iteration, target = 'contrastive_feature', smooth_weights = torch.softmax(smooth_weights, dim = -1) if smooth_weights is not None else None, smooth_type = None)
  File "/home/bxiong/workspace/LaGa/scene/__init__.py", line 240, in save_feature
    self.feature_gaussians.save_ply(os.path.join(point_cloud_path, f"{target}_point_cloud.ply"), smooth_weights, smooth_type, smooth_K, generator=generator, point_rgb=point_colors)
  File "/home/bxiong/anaconda3/envs/laga/lib/python3.10/site-packages/torch/utils/_contextlib.py", line 116, in decorate_context
    return func(*args, **kwargs)
  File "/home/bxiong/workspace/LaGa/scene/gaussian_model_ff.py", line 518, in save_ply
    f = self.get_point_features(generator = generator).detach().contiguous().cpu().numpy()
TypeError: 'Parameter' object is not callable
Training progress:  23%|███████████████▋                                                   | 7000/30000 [19:39<1:04:37,  5.93it/s, RFN=0.854, Pos cos=0.930, Neg cos=-0.029, Multi Lvl Loss=-2.228, Loss=-2.207]

How to solve?

[Jumpat]

This is a little bug brought by code version iteration. This line should be modified

f = self.get_point_features(generator = generator).detach().contiguous().cpu().numpy()

to

f = self.get_point_features.detach().contiguous().cpu().numpy()