Skip to content

slight improvements to IndexKernel for the rank == 0 (diagonal) case #2141

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
wants to merge 2 commits into
base: main
Choose a base branch
from
Open

slight improvements to IndexKernel for the rank == 0 (diagonal) case #2141

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
00a7522
slight improvements to IndexKernel for the rank = 0 (diagonal) case
rmgarnett Sep 19, 2022
a4a622b
adding rank = 0 (diagonal) branch to _eval_covar_matrix() in IndexKernel
rmgarnett Sep 20, 2022
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
20 changes: 14 additions & 6 deletions gpytorch/kernels/index_kernel.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -65,9 +65,11 @@ def __init__(
if var_constraint is None:
var_constraint = Positive()

self.register_parameter(
name="covar_factor", parameter=torch.nn.Parameter(torch.randn(*self.batch_shape, num_tasks, rank))
)
self.rank = rank
if self.rank > 0:
self.register_parameter(
name="covar_factor", parameter=torch.nn.Parameter(torch.randn(*self.batch_shape, num_tasks, self.rank))
)
self.register_parameter(name="raw_var", parameter=torch.nn.Parameter(torch.randn(*self.batch_shape, num_tasks)))
if prior is not None:
if not isinstance(prior, Prior):
Expand All @@ -88,13 +90,19 @@ def _set_var(self, value):
self.initialize(raw_var=self.raw_var_constraint.inverse_transform(value))

def _eval_covar_matrix(self):
cf = self.covar_factor
return cf @ cf.transpose(-1, -2) + torch.diag_embed(self.var)
if self.rank > 0:
cf = self.covar_factor
return cf @ cf.transpose(-1, -2) + torch.diag_embed(self.var)
else:
return torch.diag_embed(self.var)

@property
def covar_matrix(self):
var = self.var
res = PsdSumLinearOperator(RootLinearOperator(self.covar_factor), DiagLinearOperator(var))
if self.rank > 0:
res = PsdSumLinearOperator(RootLinearOperator(self.covar_factor), DiagLinearOperator(var))
else:
res = DiagLinearOperator(var)
return res

def forward(self, i1, i2, **params):
Expand Down