emmi_inference.examples.run_abupt

Functions

plot_3d_comparison(positions, target_values, ...[, ...])

Creates a 3-panel 3D scatter plot comparing target values, predicted values, and their delta.

run_abupt()

Module Contents

emmi_inference.examples.run_abupt.plot_3d_comparison(positions, target_values, predicted_values, titles=('Target', 'Predicted', 'Delta'), num_points=10000, value_clamp=None, delta_clamp=None, figsize=(18, 6), view_init=(20, 125, 0), alpha=0.5)

Creates a 3-panel 3D scatter plot comparing target values, predicted values, and their delta.

Parameters:

  • positions (torch.Tensor) – The (N, 3) tensor of 3D point coordinates.

  • target_values (torch.Tensor) – The (N,) tensor of ground truth values.

  • predicted_values (torch.Tensor) – The (N,) tensor of predicted values.

  • titles (tuple[str, str, str]) – A tuple of three strings for the subplot titles.

  • num_points (int) – Number of points to subsample for plotting.

  • value_clamp (tuple[float, float] | None) – Optional (min, max) tuple to clamp target and predicted values.

  • delta_clamp (tuple[float, float] | None) – Optional (min, max) tuple to clamp the delta values.

  • figsize (tuple[int, int]) – Figure size in inches.

  • view_init (tuple[int, int, int]) – The (elevation, azimuth, roll) camera angle for the 3D plots.

  • alpha (float) – The opacity of the scattered points.

Return type:

None

emmi_inference.examples.run_abupt.run_abupt()
Return type:

None