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] Replay buffer checkpointers #2137

Merged
merged 14 commits into from
May 16, 2024
Merged

[Feature] Replay buffer checkpointers #2137

Changes from 1 commit
Commits
Show all changes
14 commits
Select commit Hold shift + click to select a range
fce8ec7
init
vmoens Apr 29, 2024
6f9c672
amend
vmoens Apr 29, 2024
04bcce0
amend
vmoens Apr 29, 2024
2117f73
Merge remote-tracking branch 'origin/main' into register-data-hooks
vmoens Apr 30, 2024
43e227b
Merge remote-tracking branch 'origin/main' into register-data-hooks
vmoens May 3, 2024
4e86a09
t
vmoens May 7, 2024
37ca451
amend
vmoens May 7, 2024
29087c5
amend
vmoens May 7, 2024
df4280d
amend
vmoens May 8, 2024
387d14c
amend
vmoens May 8, 2024
f5282c3
Merge remote-tracking branch 'origin/main' into register-data-hooks
vmoens May 13, 2024
ccdd465
amend
vmoens May 14, 2024
94e7934
amend
vmoens May 16, 2024
2a184b3
amend
vmoens May 16, 2024
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
Next Next commit
amend
  • Loading branch information
@vmoens
vmoens committed Apr 29, 2024
commit 04bcce0f29ebf0d60c3f7293a36525385845ab5a
45 changes: 44 additions & 1 deletion torchrl/data/replay_buffers/utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -125,6 +125,23 @@ class TED2Flat:
... # load the data to represent it
... td = TensorDict.load(tmpdir + "/storage/")
... print(td)
TensorDict(
fields={
action: MemoryMappedTensor(shape=torch.Size([200, 2]), device=cpu, dtype=torch.int64, is_shared=True),
collector: TensorDict(
fields={
traj_ids: MemoryMappedTensor(shape=torch.Size([200]), device=cpu, dtype=torch.int64, is_shared=True)},
batch_size=torch.Size([]),
device=cpu,
is_shared=False),
done: MemoryMappedTensor(shape=torch.Size([200, 1]), device=cpu, dtype=torch.bool, is_shared=True),
observation: MemoryMappedTensor(shape=torch.Size([220, 4]), device=cpu, dtype=torch.float32, is_shared=True),
reward: MemoryMappedTensor(shape=torch.Size([200, 1]), device=cpu, dtype=torch.float32, is_shared=True),
terminated: MemoryMappedTensor(shape=torch.Size([200, 1]), device=cpu, dtype=torch.bool, is_shared=True),
truncated: MemoryMappedTensor(shape=torch.Size([200, 1]), device=cpu, dtype=torch.bool, is_shared=True)},
batch_size=torch.Size([]),
device=cpu,
is_shared=False)

"""

Expand Down Expand Up @@ -196,13 +213,39 @@ class Flat2TED:
... rb.dumps(tmpdir)
... # load the data to represent it
... td = TensorDict.load(tmpdir + "/storage/")
... print(td)
...
... rb_load = ReplayBuffer(storage=LazyMemmapStorage(200))
... rb_load.register_load_hook(Flat2TED())
... rb_load.load(tmpdir)
... print("storage after loading", rb_load[:])
... assert (rb[:] == rb_load[:]).all()
storage after loading TensorDict(
fields={
action: MemoryMappedTensor(shape=torch.Size([200, 2]), device=cpu, dtype=torch.int64, is_shared=False),
collector: TensorDict(
fields={
traj_ids: MemoryMappedTensor(shape=torch.Size([200]), device=cpu, dtype=torch.int64, is_shared=False)},
batch_size=torch.Size([200]),
device=cpu,
is_shared=False),
done: MemoryMappedTensor(shape=torch.Size([200, 1]), device=cpu, dtype=torch.bool, is_shared=False),
next: TensorDict(
fields={
done: MemoryMappedTensor(shape=torch.Size([200, 1]), device=cpu, dtype=torch.bool, is_shared=False),
observation: MemoryMappedTensor(shape=torch.Size([200, 4]), device=cpu, dtype=torch.float32, is_shared=False),
reward: MemoryMappedTensor(shape=torch.Size([200, 1]), device=cpu, dtype=torch.float32, is_shared=False),
terminated: MemoryMappedTensor(shape=torch.Size([200, 1]), device=cpu, dtype=torch.bool, is_shared=False),
truncated: MemoryMappedTensor(shape=torch.Size([200, 1]), device=cpu, dtype=torch.bool, is_shared=False)},
batch_size=torch.Size([200]),
device=cpu,
is_shared=False),
observation: MemoryMappedTensor(shape=torch.Size([200, 4]), device=cpu, dtype=torch.float32, is_shared=False),
terminated: MemoryMappedTensor(shape=torch.Size([200, 1]), device=cpu, dtype=torch.bool, is_shared=False),
truncated: MemoryMappedTensor(shape=torch.Size([200, 1]), device=cpu, dtype=torch.bool, is_shared=False)},
batch_size=torch.Size([200]),
device=cpu,
is_shared=False)


"""

Expand Down
Loading