parsnip.ParsnipModel.forward

ParsnipModel.forward(light_curves, sample=True, to_numpy=False)

Run a set of light curves through the full ParSNIP model

We use variational inference to predict the latent representation of each light curve, and we then use the generative model to predict the light curves for those representations.

Parameters:

  • light_curves (List[Table]) – List of light curves

  • sample (bool, optional) – If True (default), sample from the posterior distribution. If False, use the MAP.

  • to_numpy (bool, optional) – Whether to convert the outputs to numpy arrays, by default False

Returns:

Result dictionary. If to_numpy is True, all of the elements will be numpy arrays. Otherwise, they will be PyTorch tensors on the model’s device.

Return type:

dict