[BugFix] Fix invalid CUDA ID error when loading Bounded variables across devices

[cbhua]

Description

This pull request resolves an invalid CUDA ID error that occurs when transferring a Bounded variable between servers with different numbers of GPUs.

In the current implementation, when changing the device of a Bounded variable, there is:

rl/torchrl/data/tensor_specs.py

Lines 2273 to 2274 in df4fa78

	low=self.space.low.to(dest),
	high=self.space.high.to(dest),

This operation first attempts to move self.space._low to its self.space.device before transferring to the target device (dest):

rl/torchrl/data/tensor_specs.py

Lines 376 to 382 in df4fa78

	@property
	def low(self):
	return self._low.to(self.device)
	@property
	def high(self):
	return self._high.to(self.device)

This process leads to errors when a variable previously on cuda:7 (in an 8 GPU server) is loaded on a server with only one GPU, as it incorrectly attempts to access cuda:7.

Motivation and Context

The issue was identified when a model was trained and saved on a multi-GPU cluster and subsequently loaded on a local server equipped with fewer GPUs. The model’s saved state includes device information specific to the original multi-GPU environment. When attempting to assign the model to a device available on the current server, the discrepancy in device IDs between the environments leads to this bug.

This PR fixes #2420 .

• I have raised an issue to propose this change (required for new features and bug fixes)

Types of changes

• Bug fix (non-breaking change which fixes an issue)

To resolve this bug, I have adjusted the approach to device assignment. Instead of calling the saved device information from the previous cluster, we directly update the device information to match the current server’s available hardware.

In the meantime, I think there is repeat function code here:

rl/torchrl/data/tensor_specs.py

Lines 390 to 408 in df4fa78

	@low.setter
	def low(self, value):
	self.device = value.device
	self._low = value.cpu()
	@high.setter
	def high(self, value):
	self.device = value.device
	self._high = value.cpu()
	@low.setter
	def low(self, value):
	self.device = value.device
	self._low = value.cpu()
	@high.setter
	def high(self, value):
	self.device = value.device
	self._high = value.cpu()

Checklist

• I have read the CONTRIBUTION guide (required)
• I have updated the tests accordingly (required for a bug fix or a new feature).

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/rl/2421

• 📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

❌ 3 New Failures, 1 Cancelled Job, 2 Pending, 6 Unrelated Failures

As of commit 9b22619 with merge base df4fa78 ():

NEW FAILURES - The following jobs have failed:

• Habitat Tests on Linux / tests (3.9, 12.1) / linux-job (gh)
  RuntimeError: Command docker exec -t b02e37726c15b75538b4a65ae237a4b8d9f2a79f43e84eeefd250104aabff594 /exec failed with exit code 139
• Libs Tests on Linux / unittests-gym (3.9, 12.1) / linux-job (gh)
  RuntimeError: Command docker exec -t c42e0db01fb04732c9a34fc10e31a41bf467804bca7e680adc53f827eeee1baf /exec failed with exit code 1
• Lint / c-source / linux-job (gh)
  RuntimeError: Command docker exec -t 9c26df123958235ef04530198910a4255b035473462ec3d3582db72c8509fc8b /exec failed with exit code 1

CANCELLED JOB - The following job was cancelled. Please retry:

• Unit-tests on Linux / tests-olddeps (3.8, 11.6) / linux-job (gh)
  ##[error]The operation was canceled.

FLAKY - The following jobs failed but were likely due to flakiness present on trunk:

• Continuous Benchmark (PR) / CPU Pytest benchmark (gh) (detected as infra flaky with no log or failing log classifier)
• Continuous Benchmark (PR) / GPU Pytest benchmark (gh) (detected as infra flaky with no log or failing log classifier)

BROKEN TRUNK - The following jobs failed but were present on the merge base:

👉 Rebase onto the `viable/strict` branch to avoid these failures

• Build Windows Wheels / pytorch/rl (pytorch/rl, .github/scripts/td_script.sh, .github/scripts/version_script.bat, python ... / upload / wheel-py3_9-cpu (gh) (trunk failure)
• Build Windows Wheels / pytorch/rl (pytorch/rl, .github/scripts/td_script.sh, .github/scripts/version_script.bat, python ... / upload / wheel-py3_9-cuda11_8 (gh) (trunk failure)
• Build Windows Wheels / pytorch/rl (pytorch/rl, .github/scripts/td_script.sh, .github/scripts/version_script.bat, python ... / upload / wheel-py3_9-cuda12_1 (gh) (trunk failure)
• Build Windows Wheels / pytorch/rl (pytorch/rl, .github/scripts/td_script.sh, .github/scripts/version_script.bat, python ... / upload / wheel-py3_9-cuda12_4 (gh) (trunk failure)

This comment was automatically generated by Dr. CI and updates every 15 minutes.

[facebook-github-bot]

Hi @cbhua!

Thank you for your pull request and welcome to our community.

Action Required

In order to merge any pull request (code, docs, etc.), we require contributors to sign our Contributor License Agreement, and we don't seem to have one on file for you.

Process

In order for us to review and merge your suggested changes, please sign at https://code.facebook.com/cla. If you are contributing on behalf of someone else (eg your employer), the individual CLA may not be sufficient and your employer may need to sign the corporate CLA.

Once the CLA is signed, our tooling will perform checks and validations. Afterwards, the pull request will be tagged with CLA signed. The tagging process may take up to 1 hour after signing. Please give it that time before contacting us about it.

If you have received this in error or have any questions, please contact us at [email protected]. Thanks!

[vmoens]

LGTM thanks!

[vmoens]

Feel free to mark is as ready for review for me to merge it

[cbhua]

Thanks for your review! @vmoens, I saw some unpassed pytest checks, so I made this PR draft. It seems those are from GitHub action settings or dependency packages, so should be fine. I will make it ready to merge!

[vmoens]

Errors in CI are not related to this PR, happy to merge it