Adapts the gradient for the counterfactual loss function to use CoutnerfactualExplanations.jl for a model trained in R.

adapt_gradient.jl

import CounterfactualExplanations.Generators: ∂ℓ
using LinearAlgebra

# Countefactual loss:
function ∂ℓ(
    generator::AbstractGradientBasedGenerator, 
    counterfactual_state::CounterfactualState) 
  M = counterfactual_state.M
  nn = M.nn
  x′ = counterfactual_state.x′
  t = counterfactual_state.target_encoded
  R"""
  x <- torch_tensor($x′, requires_grad=TRUE)
  output <- $nn(x)
  loss_fun <- nnf_binary_cross_entropy_with_logits
  obj_loss <- loss_fun(output,$t)
  obj_loss$backward()
  """
  grad = rcopy(R"as_array(x$grad)")
  return grad
end