wuwenhai
/
bev-project
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
55 Commits 1 Branch 0 Tags 125 MiB
Commit Graph

13 Commits

Author SHA1 Message Date
bevfusion f1f37a8d79 Initial commit 2025-11-21 02:50:51 +00:00
bevfusion fcf3ae0ea9 Complete project state snapshot: Phase 4B RMT-PPAD Integration 
🎯 Training Status:
- Current Epoch: 2/10 (13.3% complete)
- Segmentation Dice: 0.9594
- Detection IoU: 0.5742
- Training stable with 8 GPUs

🔧 Technical Achievements:
- ✅ RMT-PPAD Transformer segmentation decoder integrated
- ✅ Task-specific GCA architecture optimized
- ✅ Multi-scale feature fusion (180×180, 360×360, 600×600)
- ✅ Adaptive scale weight learning implemented
- ✅ BEVFusion multi-task framework enhanced

📊 Performance Highlights:
- Divider segmentation: 0.9793 Dice (excellent)
- Pedestrian crossing: 0.9812 Dice (excellent)
- Stop line: 0.9812 Dice (excellent)
- Carpark area: 0.9802 Dice (excellent)
- Walkway: 0.9401 Dice (good)
- Drivable area: 0.8959 Dice (good)

🛠️ Code Changes Included:
- Enhanced BEVFusion model (bevfusion.py)
- RMT-PPAD integration modules (rmtppad_integration.py)
- Transformer segmentation head (enhanced_transformer.py)
- GCA module optimizations (gca.py)
- Configuration updates (Phase 4B configs)
- Training scripts and automation tools
- Comprehensive documentation and analysis reports

📅 Snapshot Date: Fri Nov 14 09:06:09 UTC 2025
📍 Environment: Docker container
🎯 Phase: RMT-PPAD Integration Complete
 2025-11-14 09:06:09 +00:00
Jeff Daily 326653dc06
[ROCm] add support for ROCm/HIP (#509) 
* [ROCm] add support for ROCm/HIP

- do not use THC.h or THH.h or THCState
- rename files to avoid collisions at build time due to hipify
- HIP host device definitions in tensorview.h
- use HIP atomicAdd() in scatter_points_cuda.cu
- add WITH_ROCM anywhere WITH_CUDA was used
- update setup.py for renamed files and ROCm/HIP torch

* enforce c++17 since pytorch requires it
 2024-07-31 00:04:51 -04:00
Kevin Shao cfc1acb42a
Fix the circular import (#466) 2023-07-23 00:15:42 -04:00
Kevin Shao d0152cf97c
Add BEVFusion-R (#443) 2023-07-07 22:53:36 -04:00
kevin 0e5b9edbc1
[debug] fix bug for nuscenes dataset (#156) 
* [debug] fix bug for nuscenes dataset

* [debug] fix bug for nuscenes dataset

[debug] fix bug for nuscenes dataset
 2022-10-10 12:12:10 -04:00
Haotian (Ken) Tang f39a4a0752
Add training details (#150) 
* [Major] Update training details.

* [Minor] Update README.md.

* [Minor] Remove comment.
 2022-09-26 18:24:38 -04:00
Haotian (Ken) Tang 36d34b73eb
[Major] Add device guard. 2022-07-11 11:05:49 -04:00
JoeyforJoy 4d8b4757a8 [fix] change the function signature of BaseTransform's forward() function, to successfully run tools/visualize.py for camera-only model 2022-06-18 21:57:40 +08:00
Haotian Tang 24502fe097 [Minor] Fix dependencies. 2022-06-16 23:23:42 +00:00
zhanjun 408b645179 Fix create data problem : missing sample_index and camera_intrinsics keys in get_data_info [#11] 2022-06-10 11:15:52 +08:00
Haotian Tang 86b077bea0 [Major] Add FLOPs counter. 2022-06-07 13:49:27 +00:00
Haotian Tang da93b27d0f [Major] Code release. 2022-06-03 04:21:18 +00:00