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

[Feature] shifted for all adv #1276

Merged
merged 2 commits into from
Jun 14, 2023
Merged

[Feature] shifted for all adv #1276

Changes from 1 commit
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
e449fdf
init
vmoens Jun 14, 2023
acd4b4f
fix
vmoens Jun 14, 2023
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
Prev Previous commit
fix
  • Loading branch information
@vmoens
vmoens committed Jun 14, 2023
commit acd4b4fb792bccb8b9874bfcdf5d85af8ad79213
10 changes: 5 additions & 5 deletions torchrl/objectives/value/advantages.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -78,7 +78,7 @@ def _call_value_nets(
ndim = i + 1
if single_call:
# get data at t and last of t+1
data = torch.cat(
data_in = torch.cat(
[
data.select(*in_keys, value_key, strict=False),
data.get("next").select(*in_keys, value_key, strict=False)[..., -1:],
Expand All @@ -91,9 +91,9 @@ def _call_value_nets(
"the value at t and t+1 cannot be retrieved in a single call without recurring to vmap when both params and next params are passed."
)
if params is not None:
value_est = value_net(data, params).get(value_key)
value_est = value_net(data_in, params).get(value_key)
else:
value_est = value_net(data).get(value_key)
value_est = value_net(data_in).get(value_key)
idx = (slice(None),) * (ndim - 1) + (slice(None, -1),)
idx_ = (slice(None),) * (ndim - 1) + (slice(1, None),)
value, value_ = value_est[idx], value_est[idx_]
Expand All @@ -116,8 +116,8 @@ def _call_value_nets(
data_out = torch.vmap(value_net, (0,))(data_in)
value_est = data_out.get(value_key)
value, value_ = value_est[0], value_est[1]
data.set(value_key, value)
data.set(("next", value_key), value_)
data.set(value_key, value)
data.set(("next", value_key), value_)
if detach_next:
value_ = value_.detach()
return value, value_
Expand Down