Skip to content

feat(distributions): support symbolic eta in LKJCholeskyCov logp #8404

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
sankhyanreyansh wants to merge 1 commit into
base: main
Choose a base branch
from
+25 −6
Open

feat(distributions): support symbolic eta in LKJCholeskyCov logp #8404

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
7 changes: 1 addition & 6 deletions pymc/distributions/multivariate.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1294,17 +1294,12 @@ def _LKJCholeksyCovRV_logp(op, values, rng, n, eta, sd_dist, **kwargs):
det_invjac = pt.log(corr_diag) - idx * pt.log(sd_vals)
det_invjac = det_invjac.sum()

# TODO: _lkj_normalizing_constant currently requires `eta` and `n` to be constants
# TODO: _lkj_normalizing_constant currently requires `n` to be a constant
try:
n = int(get_underlying_scalar_constant_value(n))
except NotScalarConstantError:
raise NotImplementedError("logp only implemented for constant `n`")

try:
eta = float(get_underlying_scalar_constant_value(eta))
except NotScalarConstantError:
raise NotImplementedError("logp only implemented for constant `eta`")

norm = _lkj_normalizing_constant(eta, n)

return norm + logp_lkj + logp_sd + det_invjac
Expand Down
24 changes: 24 additions & 0 deletions tests/distributions/test_multivariate.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -975,6 +975,30 @@ def test_no_warning_logp(self):
warnings.simplefilter("error")
m.logp()

def test_lkj_cholesky_cov_symbolic_eta(self):
with pm.Model() as model:
eta = pm.HalfNormal("eta", sigma=1.0)
sd_dist = pm.Exponential.dist(1.0)
chol = pm.LKJCholeskyCov("chol", eta=eta, n=3, sd_dist=sd_dist)

logp_fn = model.compile_logp()
dlogp_fn = model.compile_dlogp()

ip = model.initial_point()
logp_val = logp_fn(ip)
assert np.isfinite(logp_val)

dlogp_val = dlogp_fn(ip)
assert dlogp_val.size == 7
assert np.all(np.isfinite(dlogp_val))

with pm.Model() as m_invalid:
n_sym = pt.lscalar("n")
sd_dist = pm.Exponential.dist(1.0)
pm.LKJCholeskyCov("chol_invalid", eta=2.0, n=n_sym, sd_dist=sd_dist, compute_corr=False)
with pytest.raises(NotImplementedError, match="logp only implemented for constant `n`"):
m_invalid.logp()

@pytest.mark.parametrize(
"sd_dist",
[
Expand Down
Loading