l5kit.evaluation package

l5kit.evaluation.compute_mse_error_csv(ground_truth_path: str, inference_output_path: str) → numpy.ndarray
Parameters
  • ground_truth_path (str) – Path to the ground truth csv file.

  • inference_output_path (str) – Path to the csv file containing network output.

l5kit.evaluation.export_zarr_to_ground_truth_csv(zarr_dataset: l5kit.data.zarr_dataset.ChunkedStateDataset, csv_file_path: str, history_num_frames: int, future_num_frames: int, filter_agents_threshold: float, pixel_size: numpy.ndarray = array([0.25, 0.25]), history_step_size: int = 1, future_step_size: int = 1) → None

Produces a csv file containing the ground truth from a zarr file.

Parameters
  • zarr_dataset (np.ndarray) – The open zarr dataset.

  • csv_file_path (str) – File path to write a CSV to.

  • history_num_frames (int) – Amount of history frames to draw into the rasters.

  • future_num_frames (int) – Amount of history frames to draw into the rasters.

  • filter_agents_threshold (float) – Value between 0 and 1 to use as cutoff value for agent filtering

  • pixel_size (np.ndarray) –

  • history_step_size (int) – Steps to take between frames, can be used to subsample history frames.

  • future_step_size (int) – Steps to take between targets into the future.

  • on their probability of being a relevant agent. (based) –

l5kit.evaluation.write_coords_as_csv(csv_file_path: str, future_num_frames: int, future_coords_offsets: numpy.ndarray, timestamps: numpy.ndarray, agent_ids: numpy.ndarray) → None

Write coordinates as a csv file

Parameters
  • csv_file_path (str) – csv path

  • future_num_frames (int) – numbers of frames in the future for each prediction

  • future_coords_offsets (np.ndarray) – array of size N x future_num_frames of displacements

  • timestamps (np.ndarray) – array of size N with timestamps int64

  • agent_ids (np.ndarray) – array for size N with track_ids int64

Returns: