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

[BugFix] Allow for composite action distributions in PPO/A2C losses #2391

Merged
merged 19 commits into from
Sep 4, 2024
Merged

[BugFix] Allow for composite action distributions in PPO/A2C losses #2391

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
19 commits
Select commit Hold shift + click to select a range
627b673
account for composite distribution
albertbou92 Aug 10, 2024
9b08555
tests
albertbou92 Aug 11, 2024
a668d0c
fix tests ppo
albertbou92 Aug 12, 2024
e65e1a3
fix tests ppo
albertbou92 Aug 12, 2024
fd60ad7
fix tests ppo
albertbou92 Aug 12, 2024
bc00e2d
fix tests ppo
albertbou92 Aug 12, 2024
e3f61f7
fix tests ppo
albertbou92 Aug 12, 2024
4eb413d
fix tests ppo
albertbou92 Aug 12, 2024
560ddad
fix tests ppo
albertbou92 Aug 12, 2024
2865c92
format
albertbou92 Aug 12, 2024
13a092d
a2c tests
albertbou92 Aug 14, 2024
8bc8e03
a2c tests
albertbou92 Aug 14, 2024
07ec262
a2c tests
albertbou92 Aug 14, 2024
a95eb3d
a2c tests
albertbou92 Aug 14, 2024
8f00828
format
albertbou92 Aug 14, 2024
9ce3c58
docstrings and minor fixes
albertbou92 Aug 14, 2024
ad32b90
docstrings and minor fixes
albertbou92 Aug 14, 2024
37b733f
docstrings fix
albertbou92 Aug 14, 2024
69922fa
added required grad for td
albertbou92 Aug 26, 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
docstrings and minor fixes
  • Loading branch information
@albertbou92
albertbou92 committed Aug 14, 2024
commit 9ce3c5876b977f1dc004979b929e2ad908a73098
19 changes: 16 additions & 3 deletions test/test_cost.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9238,13 +9238,26 @@ def test_a2c_tensordict_keys_run(
@pytest.mark.parametrize("reward_key", ["reward", "reward2"])
@pytest.mark.parametrize("done_key", ["done", "done2"])
@pytest.mark.parametrize("terminated_key", ["terminated", "terminated2"])
@pytest.mark.parametrize("composite_action_dist", [False, ])
@pytest.mark.parametrize(
"composite_action_dist",
[
False,
],
)
def test_a2c_notensordict(
self, action_key, observation_key, reward_key, done_key, terminated_key, composite_action_dist
self,
action_key,
observation_key,
reward_key,
done_key,
terminated_key,
composite_action_dist,
):
torch.manual_seed(self.seed)

actor = self._create_mock_actor(observation_key=observation_key, composite_action_dist=composite_action_dist)
actor = self._create_mock_actor(
observation_key=observation_key, composite_action_dist=composite_action_dist
)
value = self._create_mock_value(observation_key=observation_key)
td = self._create_seq_mock_data_a2c(
action_key=action_key,
Expand Down
16 changes: 14 additions & 2 deletions torchrl/objectives/a2c.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -10,7 +10,12 @@
from typing import Tuple

import torch
from tensordict import TensorDict, TensorDictBase, TensorDictParams
from tensordict import (
is_tensor_collection,
TensorDict,
TensorDictBase,
TensorDictParams,
)
from tensordict.nn import dispatch, ProbabilisticTensorDictSequential, TensorDictModule
from tensordict.utils import NestedKey
from torch import distributions as d
Expand Down Expand Up @@ -190,6 +195,13 @@ class A2CLoss(LossModule):
... next_reward = torch.randn(*batch, 1),
... next_observation = torch.randn(*batch, n_obs))
>>> loss_obj.backward()

.. note::
There is an exception regarding compatibility with non-tensordict-based modules.
If the actor network is probabilistic and uses a `~tensordict.nn.distributions.CompositeDistribution`,
this class must be used with tensordicts and cannot function as a tensordict-independent module.
This is because composite action spaces inherently rely on the structured representation of data provided by
tensordicts to handle their actions.
"""

@dataclass
Expand Down Expand Up @@ -384,7 +396,7 @@ def get_entropy_bonus(self, dist: d.Distribution) -> torch.Tensor:
except NotImplementedError:
x = dist.rsample((self.samples_mc_entropy,))
log_prob = dist.log_prob(x)
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Previously, was this a bug or did we sum the log-probs automatically?

Copy link
Contributor Author

@albertbou92 albertbou92 Aug 26, 2024 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This is simply because the log_prob() method for a composite dist will return a TD instead of a Tensor, so we compute the entropy in 2 steps.

This is the old version:

    def get_entropy_bonus(self, dist: d.Distribution) -> torch.Tensor:
        try:
            entropy = dist.entropy()
        except NotImplementedError:
            x = dist.rsample((self.samples_mc_entropy,))
            entropy = -dist.log_prob(x).mean(0)
        return entropy.unsqueeze(-1)

This is the new version. It simply retrieves the log tensor before computing the entropy.

    def get_entropy_bonus(self, dist: d.Distribution) -> torch.Tensor:
        try:
            entropy = dist.entropy()
        except NotImplementedError:
            x = dist.rsample((self.samples_mc_entropy,))
            log_prob = dist.log_prob(x)
            if is_tensor_collection(log_prob):
                log_prob = log_prob.get(self.tensor_keys.sample_log_prob)
            entropy = -log_prob.mean(0)
        return entropy.unsqueeze(-1)

if isinstance(log_prob, TensorDict):
if is_tensor_collection(log_prob):
log_prob = log_prob.get(self.tensor_keys.sample_log_prob)
entropy = -log_prob.mean(0)
return entropy.unsqueeze(-1)
Expand Down
17 changes: 14 additions & 3 deletions torchrl/objectives/ppo.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -12,7 +12,12 @@
from typing import Tuple

import torch
from tensordict import TensorDict, TensorDictBase, TensorDictParams
from tensordict import (
is_tensor_collection,
TensorDict,
TensorDictBase,
TensorDictParams,
)
from tensordict.nn import (
dispatch,
ProbabilisticTensorDictModule,
Expand Down Expand Up @@ -238,6 +243,12 @@ class PPOLoss(LossModule):
... next_observation=torch.randn(*batch, n_obs))
>>> loss_objective.backward()

.. note::
There is an exception regarding compatibility with non-tensordict-based modules.
If the actor network is probabilistic and uses a `~tensordict.nn.distributions.CompositeDistribution`,
this class must be used with tensordicts and cannot function as a tensordict-independent module.
This is because composite action spaces inherently rely on the structured representation of data provided by
tensordicts to handle their actions.
"""

@dataclass
Expand Down Expand Up @@ -450,7 +461,7 @@ def get_entropy_bonus(self, dist: d.Distribution) -> torch.Tensor:
except NotImplementedError:
x = dist.rsample((self.samples_mc_entropy,))
log_prob = dist.log_prob(x)
if isinstance(log_prob, TensorDict):
if is_tensor_collection(log_prob):
log_prob = log_prob.get(self.tensor_keys.sample_log_prob)
entropy = -log_prob.mean(0)
return entropy.unsqueeze(-1)
Expand Down Expand Up @@ -1119,7 +1130,7 @@ def forward(self, tensordict: TensorDictBase) -> TensorDict:
x = previous_dist.sample((self.samples_mc_kl,))
previous_log_prob = previous_dist.log_prob(x)
current_log_prob = current_dist.log_prob(x)
if isinstance(x, TensorDict):
if is_tensor_collection(x):
previous_log_prob = previous_log_prob.get(
self.tensor_keys.sample_log_prob
)
Expand Down