helper functions

hopfield4py.hamming(A: Tensor, B: Tensor) → Tensor

Hamming distance sum(abs((A-B)))/lenght(A)

A and be must have the same shape

Parameters:

  • A – first tensor

  • B – second tensor

hopfield4py.get_prediction(sample: Tensor, data_tensor: Tensor, model: Hopfield) → Tensor

Get the nearest memory

Parameters:

  • sample – sample to reconstruct

  • data_tensor – tensor with memories

  • model – model used to infer

Returns:

argmin of the element of data_tensor nearest to sample

hopfield4py.get_predicted_labels(classes: list, samples: Tensor, data_tensor: Tensor, model: Hopfield) → list

Get the classes predicted for each sample

Parameters:

  • classes – list of classes names with shape (nclasses,)

  • samples – samples to reconstruct with shape (nsamples, nspins)

  • data_tensor – tensor with memories (nclasses, nspins)

Returns:

tensor with list of classes

hopfield4py.get_all_prediction(samples: Tensor, data_tensor: Tensor, model: Hopfield) → Tensor

Get the nearest memory

Parameters:

  • samples – samples to reconstruct

  • data_tensor – tensor with memories

  • model – model used to infer

Returns:

tensor with list of argmin of the element of data_tensor nearest to each sample