Fix hardcoded CUDA device in api.py to support MPS and CPU fallback#1516
Open
Mr-Neutr0n wants to merge 1 commit intozai-org:mainfrom
Open
Fix hardcoded CUDA device in api.py to support MPS and CPU fallback#1516Mr-Neutr0n wants to merge 1 commit intozai-org:mainfrom
Mr-Neutr0n wants to merge 1 commit intozai-org:mainfrom
Conversation
The API server crashes on non-CUDA systems (e.g., macOS with Apple Silicon or CPU-only machines) because DEVICE is hardcoded to "cuda" and the model loading uses .cuda() directly. Changes: - Auto-detect the best available device (CUDA > MPS > CPU) - Replace .cuda() with .to(DEVICE) for portable device placement - Update torch_gc() to handle MPS cache clearing and skip CUDA-specific cleanup when not on a CUDA device
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Problem
The API server (
api.py) crashes immediately on non-CUDA systems because:DEVICEis hardcoded to"cuda", causing failures on macOS (Apple Silicon/MPS) and CPU-only machines.cuda()directly, which raisesRuntimeErrorwhen CUDA is unavailabletorch_gc()only handles CUDA cleanup, missing MPS cache managementFix
.cuda()with.to(DEVICE)for portable device placementtorch_gc()to clear MPS cache when on Apple Silicon and skip CUDA-specific cleanup when not on a CUDA deviceTesting
torch.cuda.is_available()returnsTrueand DEVICE resolves to"cuda"as before