NeuralTracking icon indicating copy to clipboard operation
NeuralTracking copied to clipboard

Many bugs in the code.

Open ootts opened this issue 2 years ago • 2 comments

Hi, I find your code very helpful in reproducing NNRT+Fusion. However, when I run "python run_fusion.py", there are lots of errors raised. I am wondering if I am running the correct version of your code. I am using the main branch, the latest commit. The errors include but not limited:

  1. Parameter has no attribute index. Currently I change the .index to .value.
  2. graph_for_integration is not defined.
  3. here should be "color".
  4. here warped_mesh is not defined.

Please kindly guide me to run it successfully! Thanks!

ootts avatar Jan 24 '24 13:01 ootts

Hello @ootts , I've also been working with this repo.

Another issue I met while running "run_fusion.py" is that the "apps/create_graph_data.py" calls the GraphWarpField() function. In python codes, the function takes 4 parameters: (nodes, edges, edge_weights, clusters).

However, currently the GraphWarpField() function is binded to "HierarchicalGraphWarpField.h" that only takes nodes in it's constructor.

Thus, whenever GraphWarpField() is called in python code, an error is thrown.

I'm guessing that it is supposed to be binded to "PlanarGraphWarpFields.h", which takes (nodes, edges, edge_weights, clusters).

Please let me know if you have resolved this issue

Leeffkk avatar Jan 29 '24 05:01 Leeffkk

@Leeffkk Hi, I also met this issue. I skipped this issue by returning the edges, edge_weights, and clusters separately. The GraphWarpField class is kept as only taking nodes as member variables.

ootts avatar Jan 29 '24 13:01 ootts