Fix CPU support by removing hardcoded CUDA calls - #81
Open
jonathansim94 wants to merge 1 commit into
Open
Conversation
meta-cla
Bot
added
the
CLA Signed
|
This is also an issue with multi-gpu setups. Hardcoded |
adcoh
added a commit
to adcoh/sam-3d-body
that referenced
this pull request
Apr 8, 2026
Fixes facebookresearch#93. Supersedes facebookresearch#81. - Replace hardcoded "cuda" / .cuda() calls with self.device - Auto-detect device: CUDA > MPS > CPU - Force MHR TorchScript to load/run on CPU (MPS incompatible with float64 placeholder storage) - Downcast float64 to float32 in recursive_to() for MPS compatibility Tested on MacBook Pro M1 Max, macOS 15.x, PyTorch 2.5+ with MPS backend.
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
This PR fixes CPU support by removing all hard-coded CUDA device calls in the model execution path. Several modules previously forced tensors to "cuda" (e.g., via .cuda(), recursive_to(..., "cuda"), or device="cuda"), which caused runtime failures when no GPU was available.
All these occurrences have been replaced with device-agnostic logic relying on "self.device".
Also tested in a Docker container without GPU support, confirming that inference runs successfully on CPU without errors.