Issues

• [Tip] Running BEVFormer on PyTorch 2.x + Blackwell (B200/H200) — Docker image + a small DDP fix

  fundamentalvision/BEVFormer#321 · Filter by author nabe2030 opened on May 22, 2026

• 是不是没有道路分割只有检测啊？no road segment

  fundamentalvision/BEVFormer#319 · Filter by author GinKenneth opened on Jan 23, 2026

• 代码里空间注意力特征聚合好像有问题

  fundamentalvision/BEVFormer#318 · Filter by author 15900854096 opened on Jan 9, 2026

• 请问是否可以使用bevformer进行3D占用预测？

  fundamentalvision/BEVFormer#317 · Filter by author suayu opened on Dec 15, 2025

• where the code for the multi-modal BEVFormer

  fundamentalvision/BEVFormer#316 · Filter by author Q1ng-mu opened on Dec 15, 2025
  4

• which is prediction or GT

  fundamentalvision/BEVFormer#315 · Filter by author icefrogplus opened on Dec 1, 2025

• a little question

  fundamentalvision/BEVFormer#314 · Filter by author YWliu040120 opened on Nov 24, 2025

• mmcv version incompatible

  fundamentalvision/BEVFormer#313 · Filter by author yabuke opened on Oct 20, 2025
  2

• Difficulty of Running BEV Former on Blackwell Architecture GPU RTX5060, Request for Update & solution.

  fundamentalvision/BEVFormer#311 · Filter by author ThuraTunScibotics opened on Sep 28, 2025
  7

• Question about rotation_angle when rotating prev_bev

  fundamentalvision/BEVFormer#310 · Filter by author whystopped opened on Sep 24, 2025

• TemporalSelfAttention在使用pre_bev去拼接query时是否存在问题？

  fundamentalvision/BEVFormer#309 · Filter by author hxdoit opened on Aug 14, 2025
  4

• Cannot find the code for the multi-modal BEVformer

  fundamentalvision/BEVFormer#308 · Filter by author Q1ng-mu opened on Jul 29, 2025
  2

• Transform the coordinates of a real object to BEV representation

  fundamentalvision/BEVFormer#306 · Filter by author tjyuyao opened on Jul 4, 2025

• 运行安装命令时报错：“ModuleNotFoundError: No module named 'commands'”，同时安装过程中发现多处warning如warning: no files found matching 'mmdet3d/.mim/model-index.yml'

  fundamentalvision/BEVFormer#304 · Filter by author xli-li opened on May 19, 2025
  1

• don't use load_from = 'ckpts/r101_dcn_fcos3d_pretrain.pth'

  fundamentalvision/BEVFormer#301 · Filter by author HS-FFAD opened on Apr 23, 2025

• Why rotate only the ground truth

  fundamentalvision/BEVFormer#300 · Filter by author tianjunk opened on Apr 8, 2025

• Where is Bevformer-M in your latest Tpami paper

  fundamentalvision/BEVFormer#299 · Filter by author hangzheng-yu opened on Mar 17, 2025
  2

• completely remove can-bus from the project

  fundamentalvision/BEVFormer#298 · Filter by author xudh777 opened on Mar 12, 2025
  3

• Can you help me with a question? Thank you

  fundamentalvision/BEVFormer#297 · Filter by author uaing opened on Mar 10, 2025
  1

• Decoder with Hybrid Object Queries

  fundamentalvision/BEVFormer#296 · Filter by author Love-syntacticSugar opened on Feb 27, 2025

• error: the following arguments are required: config

  fundamentalvision/BEVFormer#295 · Filter by author woshini-baba opened on Feb 21, 2025

• any update on segmentation model ?

  fundamentalvision/BEVFormer#290 · Filter by author Rushikesh22750 opened on Dec 20, 2024

• Help....How to evaluate on pretrained models ??

  fundamentalvision/BEVFormer#289 · Filter by author Rushikesh22750 opened on Dec 16, 2024

• 请问如何获取模型的参数量和计算量

  fundamentalvision/BEVFormer#288 · Filter by author gaoxuezhao opened on Dec 6, 2024

• 大佬们，我想只取六个相机的两个视角，怎么改配置文件啊

  fundamentalvision/BEVFormer#287 · Filter by author fangwt-1 opened on Dec 2, 2024
  2