Fixes default_dqn_torch_rl_module assuming the device is 'cpu' #54004
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
Previously, this code would produce 'RuntimeError: Expected all tensors to be on the same device, but found at least two devices, cuda:0 and cpu!' when the module was being run on GPU Signed-off-by: Max Ward <maxhwardg@users.noreply.github.com>
There was a problem hiding this comment.
Pull Request Overview
This PR fixes the device mismatch issue in the DQN module by ensuring that operations are performed on the same device.
- Adds a pre-condition check that raises an error if random_actions and exploit_actions are on different devices.
- Converts the output of torch.rand() to the device of exploit_actions for consistency.
Comments suppressed due to low confidence (1)
rllib/algorithms/dqn/torch/default_dqn_torch_rl_module.py:103
- [nitpick] The explicit device conversion reinforces device consistency; verify that this matches the intended design of always aligning random_actions with exploit_actions.
torch.rand((B,)).to(exploit_actions.device) < epsilon,
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com> Signed-off-by: Max Ward <maxhwardg@users.noreply.github.com>
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Outdated suggestions cannot be applied.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.
Previously, this code would produce 'RuntimeError: Expected all tensors to be on the same device, but found at least two devices, cuda:0 and cpu!' when the module was being run on GPU.
Why are these changes needed?
They're breaking the DQN algorithm when using the GPU to run inference.
Related issue number
N/A
Checks
git commit -s) in this PR.scripts/format.shto lint the changes in this PR.method in Tune, I've added it in
doc/source/tune/api/under thecorresponding
.rstfile.