fix(pu): fix noise layer's usage #866
Open
+97
−5
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
PaParaZz1
requested changes
Apr 18, 2025
| @@ -57,6 +58,8 @@ def __init__( | |||
| - dropout (:obj:`Optional[float]`): The dropout rate of the dropout layer. \ | |||
| if ``None`` then default disable dropout layer. | |||
| - init_bias (:obj:`Optional[float]`): The initial value of the last layer bias in the head network. \ | |||
| - noise (:obj:`bool`): Whether use ``NoiseLinearLayer`` as ``layer_fn`` in Q networks' MLP. \ | |||
There was a problem hiding this comment.
- to use
- use
NoiseLinearLayerto boost exploration
| @@ -57,6 +58,8 @@ def __init__( | |||
| - dropout (:obj:`Optional[float]`): The dropout rate of the dropout layer. \ | |||
| if ``None`` then default disable dropout layer. | |||
| - init_bias (:obj:`Optional[float]`): The initial value of the last layer bias in the head network. \ | |||
| - noise (:obj:`bool`): Whether use ``NoiseLinearLayer`` as ``layer_fn`` in Q networks' MLP. \ | |||
| Default ``False``. | |||
| @@ -384,6 +386,12 @@ def _forward_collect(self, data: Dict[int, Any], eps: float) -> Dict[int, Any]: | |||
| data = default_collate(list(data.values())) | |||
| if self._cuda: | |||
| data = to_device(data, self._device) | |||
| # Use the add_noise parameter to decide noise mode. | |||
There was a problem hiding this comment.
rename to noisy_net field and add it into default config (in the policy level), don't use xxx.get
| @@ -248,6 +248,8 @@ def _forward_learn(self, data: List[Dict[str, Any]]) -> Dict[str, Any]: | |||
| .. note:: | |||
| For more detailed examples, please refer to our unittest for DQNPolicy: ``ding.policy.tests.test_dqn``. | |||
| """ | |||
| set_noise_mode(self._learn_model, True) | |||
There was a problem hiding this comment.
use noisy_net to control this line
Another question: how to deal with target_model in noisy net
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
This PR fixes [Issue #850](#850) by adding configurable support for NoisyNet in collect mode for DQN. It introduces a new
force_noiseflag inNoiseLinearLayerand a helper functionset_noise_mode. A new config parameter (collect.add_noise) is added so that if set to True during collection, noise is injected; otherwise, it is disabled.Related Issue
TODO
force_noisecontrol inNoiseLinearLayerset_noise_modehelper to update noise settings_forward_collectand_forward_evalmethods based oncollect.add_noiseCheck List