Skip to content
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

Jump to bottom

flatten pB action dims for parameter learning with multi actions #161

Merged
merged 1 commit into from
Dec 2, 2024
Merged

flatten pB action dims for parameter learning with multi actions #161

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
20 changes: 11 additions & 9 deletions pymdp/agent.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -180,7 +180,7 @@ def __init__(
policy_len,
control_fac_idx,
)
B, self.action_maps = self._flatten_B_action_dims(B, self.B_action_dependencies)
B, pB, self.action_maps = self._flatten_B_action_dims(B, pB, self.B_action_dependencies)
policies = self._construct_flattend_policies(policies_multi, self.action_maps)
self.sampling_mode = "full"

Expand Down Expand Up @@ -575,26 +575,26 @@ def _construct_dependencies(self, A_dependencies, B_dependencies, B_action_depen
B_action_dependencies = [[f] for f in range(self.num_factors)]
return A_dependencies, B_dependencies, B_action_dependencies

def _flatten_B_action_dims(self, B, B_action_dependencies):
def _flatten_B_action_dims(self, B, pB, B_action_dependencies):
assert hasattr(B[0], "shape"), "Elements of B must be tensors and have attribute shape"
action_maps = [] # mapping from multi action dependencies to flat action dependencies for each B
B_flat = []
pB_flat = []
for i, (B_f, action_dependency) in enumerate(zip(B, B_action_dependencies)):
if action_dependency == []:
B_flat.append(jnp.expand_dims(B_f, axis=-1))
if pB is not None:
pB_flat.append(jnp.expand_dims(pB[i], axis=-1))
action_maps.append(
{
"multi_dependency": [],
"multi_dims": [],
"flat_dependency": [i],
"flat_dims": [1],
}
{"multi_dependency": [], "multi_dims": [], "flat_dependency": [i], "flat_dims": [1]}
)
continue

dims = [self.num_controls_multi[d] for d in action_dependency]
target_shape = list(B_f.shape)[: -len(action_dependency)] + [pymath.prod(dims)]
B_flat.append(B_f.reshape(target_shape))
if pB is not None:
pB_flat.append(pB[i].reshape(target_shape))
action_maps.append(
{
"multi_dependency": action_dependency,
Expand All @@ -603,7 +603,9 @@ def _flatten_B_action_dims(self, B, B_action_dependencies):
"flat_dims": [pymath.prod(dims)],
}
)
return B_flat, action_maps
if pB is None:
pB_flat = None
return B_flat, pB_flat, action_maps

def _construct_flattend_policies(self, policies, action_maps):
policies_flat = []
Expand Down