feat: Edge deployment optimization with GPU-native BEV, modern CUDA ops, and gradient accumulation #12
Open
fthbng77 wants to merge 5 commits intoTRV-Lab:masterfrom
Open
feat: Edge deployment optimization with GPU-native BEV, modern CUDA ops, and gradient accumulation #12fthbng77 wants to merge 5 commits intoTRV-Lab:masterfrom
fthbng77 wants to merge 5 commits intoTRV-Lab:masterfrom
Conversation
- Replace deprecated THC/THC.h with ATen/cuda/CUDAContext.h in all CUDA extension cpp files - Remove obsolete extern THCState declarations - Update VoD dataset paths to match local directory structure - Add .gitignore for build artifacts, data, and IDE files - Add anchor visualization tool Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Reduce BATCH_SIZE_PER_GPU to 1 with ACCUMULATION_STEPS=4 for 8GB VRAM - Implement gradient accumulation in train_one_epoch - Remove compiled .so files from git (already in .gitignore)
…fa's CPU-based DBSCAN for edge optimization, updating the image backbone configuration and documentation.
…ataset, leveraging GridDensityBEV, AMP, and a dedicated configuration.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
This PR introduces several improvements aimed at edge deployment optimization
and modern PyTorch compatibility:
clustering with a GPU-native grid density approach for edge deployment scenarios
THC/THC.hheaders withATen/cuda/CUDAContext.hacross all CUDA extension files, removing obsoleteextern THCStatedeclarations for PyTorch 2.x support(8GB VRAM) by reducing batch size to 1 with 4 accumulation steps
.sobinaries from tracking (should be builtfrom source per environment)
Changes
pcdet/models/backbones_image/CQCA_cfa.py— New GridDensityBEV modulepcdet/ops/**/src/*.cpp— Updated CUDA extension headers for PyTorch 2.xtools/train_utils/train_utils.py— Gradient accumulation in training looptools/cfgs/MAFF-Net/MAFF-Net_vod.yaml— Config updates for edge deployment.sofiles from version controlMotivation