Domino multiscale geometry features capability and inference script #788
Conversation
| @@ -50,7 +50,7 @@ variables: | |||
| # The following is for AWS DrivAer dataset. | |||
| UMeanTrim: vector | |||
| pMeanTrim: scalar | |||
| nutMeanTrim: scalar | |||
| # nutMeanTrim: scalar | |||
There was a problem hiding this comment.
Just checking: this is supposed to be commented out, right?
| @@ -233,6 +233,15 @@ def __getitem__(self, idx): | |||
| volume_coordinates = volume_coordinates[ids_in_bbox] | |||
| volume_fields = volume_fields[ids_in_bbox] | |||
|
|
|||
| # volume_fields[:, :3] = volume_fields[:, :3] / STREAM_VELOCITY | |||
There was a problem hiding this comment.
Do you want to keep it commented out or remove? If the former - maybe add a comment/TODO.
|
|
||
| self.bq_warp = nn.ModuleList() | ||
| self.geo_processors = nn.ModuleList() | ||
| for j, p in enumerate(radii): |
There was a problem hiding this comment.
Nit: can use simply for r in radii
| model_parameters=geometry_rep.geo_processor, | ||
| ) | ||
| self.geo_conv_out = nn.ModuleList() | ||
| for j, p in enumerate(radii): |
There was a problem hiding this comment.
Same here - j and p are not used here at all. Can do something like:
for _ in range(len(radii)):or just fold this loop into the previous one.
| ) | ||
| self.radius = model_parameters.geometry_local.radii | ||
| self.bq_warp = nn.ModuleList() | ||
| for ct, j in enumerate(self.radius): |
There was a problem hiding this comment.
- What is
ct? jlooks like an index, but it's not - it's actual value of a radius.
Maybe refactor it for clarity?
|
|
||
| surf_max = s_max + (s_max - s_min) / 2 | ||
| surf_min = s_min - (s_max - s_min) / 2 | ||
| surf_min[2] = s_min[2] |
There was a problem hiding this comment.
What does it do? Is it z coordinate?
| surf_sdf_grid = torch.unsqueeze(surf_sdf_grid, 0) | ||
| max_min = [c_min, c_max] | ||
| surf_max_min = [surf_min, surf_max] | ||
| center_of_mass = center_of_mass |
There was a problem hiding this comment.
Does not seem to be doing anything?
| surf_max_min = [surf_min, surf_max] | ||
| center_of_mass = center_of_mass | ||
|
|
||
| return ( |
There was a problem hiding this comment.
Nit: I would try not return 8-element tuple here. Maybe use dict or named tuple or dataclass.
| ) | ||
|
|
||
|
|
||
| class dominoInference: |
There was a problem hiding this comment.
Use PascalCase for class names.
|
|
||
| domino = dominoInference(cfg, dist, False) | ||
| domino.initialize_model( | ||
| model_path="/lustre/rranade/domino_automotive_aero_nim/automotive-aerodynamics-nim/nims/domino-automotive-aero/factory/triton/model/1/checkpoints/DoMINO.0.618.pt" |
There was a problem hiding this comment.
Use config, do not hardcode.
|
/blossom-ci |
1 similar comment
|
/blossom-ci |
Modulus Pull Request
Description
This PR adds a feature in domino to encode geometry using multi-scale features. It also provides an inference script to evaluate domino on unknown STLs and flow speeds.
Checklist
Dependencies