Skip to content

Instantly share code, notes, and snippets.

@bbbales2
Last active September 23, 2020 23:55
Show Gist options
  • Star 0 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
  • Save bbbales2/a1b2bcc1f616820ce5db59b9be46c49c to your computer and use it in GitHub Desktop.
Save bbbales2/a1b2bcc1f616820ce5db59b9be46c49c to your computer and use it in GitHub Desktop.
List of functions in stan/math/rev/fun that need reverse mode implementations for var_type<Eigen::MatrixXd>
adj_jac_apply - deprecated
# Fully converted to var<mat>
softmax
dot_self
columns_dot_self
determinant
inverse
sum
multiply
log_determinant
multiply_lower_tri_self_transpose
tcrossprod
matrix_power
# Mid conversion
log_softmax
log_sum_exp
log_determinant_spd
ordered_constrain
# Converted to tadj, not yet var<mat>
columns_dot_product - done
cov_exp_quad - done
divide - done
dot_product - done
gp_exp_quad_cov - done
gp_periodic_cov - done
matrix_exp_multiply - no reverse mode specialization
mdivide_left - done
mdivide_left_spd - done
rows_dot_prod - done
mdivide_left_tri - done
quad_form - done
trace_quad_form - done
trace_gen_quad_form - done
mdivide_left_ldlt
trace_gen_inv_quad_form_ldlt
trace_inv_quad_form_ldlt
positive_ordered_constrain - done
simplex_constrain - done
unit_vector_constrain - done
sd - done
variance - done
squared_distance - done
log_determinant_ldlt
log_softmax - done
log_sum_exp - done
# Not sure there is anything to do
to_var
value_of
value_of_rec
grad
initialize_variable
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment