mit-han-lab/bevfusion issues and pull requests

#631 - Problems about tasks, min_radius, nms_type and nms_scale

Issue - State: closed - Opened by zhujiagang 4 months ago - 1 comment

#630 - Inconsistent reproducing results on the nuscene dataset!

Issue - State: closed - Opened by zyqww 4 months ago - 2 comments

#629 - Inconsistent reproducing results on the nuscene dataset!

Issue - State: closed - Opened by zyqww 4 months ago

#628 - KeyError: 'radar'

Issue - State: closed - Opened by F0sakens3-1415 4 months ago - 1 comment

#627 - How to make custom dataset only with Lidar information？

Issue - State: closed - Opened by mx2013713828 5 months ago - 1 comment

#626 - Results for 1 stage

Issue - State: closed - Opened by konyul 5 months ago - 1 comment

#625 - Clarification on Training of 'swint-nuimages-pretrained.pth'

Issue - State: closed - Opened by Ruvennsiow 5 months ago - 1 comment

#624 - Consultation on Windows training file modification

Issue - State: closed - Opened by huidabao 5 months ago - 1 comment

#623 - What changes are needed to modify the detection range to 100 meters before and after？

Issue - State: closed - Opened by Tamako668 5 months ago - 1 comment

#622 - About the test accuracy not reaching

Issue - State: closed - Opened by npuxdy 5 months ago - 2 comments

#621 - config and training setps for train a multitask model?

Issue - State: closed - Opened by shupinghu 6 months ago - 2 comments

#620 - Visualization from pkl

Issue - State: closed - Opened by gerardmartin2 6 months ago

#619 - mAP The difference between train and test AND what is the training strategy

Issue - State: closed - Opened by Li-Whasaka 6 months ago - 7 comments

#618 - config for camera+lidar with centerhead

Issue - State: closed - Opened by zyh321 6 months ago - 1 comment

#617 - Extrinsic and intrinsic matrices after ImageAug3D and GlobalRotScaleTrans

Issue - State: closed - Opened by ME-SJTU 6 months ago - 1 comment

#616 - Input/output shapes of the BEVFusion.voxelize() method

Issue - State: closed - Opened by maxvandenhoven 6 months ago - 1 comment

#615 - RuntimeError: CUDA error: CUBLAS_STATUS_ARCH_MISMATCH

Issue - State: closed - Opened by shiliashili 6 months ago - 1 comment

#614 - I met errors,please help

Issue - State: closed - Opened by sunpian1 6 months ago - 1 comment

#613 - How to change the lidar backbone to pointpillars

Issue - State: closed - Opened by wyf0414 7 months ago - 6 comments

#611 - what is the lidar-only-det.pth

Issue - State: closed - Opened by wyf0414 7 months ago - 2 comments

#607 - The problem with the test

Issue - State: closed - Opened by Ange1ika 7 months ago - 10 comments

#604 - problems about test and visualize

Issue - State: closed - Opened by 18lalala 8 months ago - 1 comment

#602 - RuntimeError: CUDA error: CUBLAS_STATUS_INVALID_VALUE when calling `cublasGemmEx( handle, opa, opb, m, n, k, &falpha, a, CUDA_R_16F, lda, b, CUDA_R_16F, ldb, &fbeta, c, CUDA_R_16F, ldc, CUDA_R_32F, CUBLAS_GEMM_DFALT_TENSOR_OP)`

Issue - State: closed - Opened by lingxe01 8 months ago - 2 comments

#600 - Why is the height of LiDAR voxelized grid 41

Issue - State: closed - Opened by duanmushuangquan 9 months ago - 1 comment

#599 - The training of bevfusion only modifies the POINT-CLOUD-RANGE algorithm, resulting in significant differences in effectiveness

Issue - State: closed - Opened by zzwforever 9 months ago - 1 comment

#597 - KeyError: "BEVFusion: 'encoders.camera.backbone.stages.0.blocks.0.attn.w_msa.relative_position_bias_table'"

Issue - State: closed - Opened by fdy61 9 months ago - 12 comments

#596 - Error occurrence in evaluating pretrained models，about the bevfusion-det.pth.

Issue - State: closed - Opened by fdy61 9 months ago - 22 comments

#594 - CAN'T REPRODUCE THE RESULT IN PAPER

Issue - State: closed - Opened by Li-Whasaka 10 months ago - 2 comments

#593 - error raised by setup.py

Issue - State: closed - Opened by fdy61 10 months ago - 1 comment

#591 - Question about how to pretrain swin-T on nuimages?

Issue - State: closed - Opened by liyih 10 months ago - 1 comment

#590 - learning rate configuration

Issue - State: closed - Opened by konyul 10 months ago - 3 comments

#586 - Missing some implementation of transfusion?

Issue - State: closed - Opened by YueSun0609 10 months ago - 1 comment

#583 - Evaluation Process is stuck vol.2

Issue - State: closed - Opened by EpicGilgamesh 10 months ago - 1 comment

#582 - Performance is worse while training

Issue - State: closed - Opened by 186big 10 months ago - 6 comments

#581 - performance drops - float16 vs float32

Issue - State: closed - Opened by tomschw800 11 months ago - 1 comment

#580 - Key Error: 'ann_info' when evaluating on Nuscenes test set

Issue - State: closed - Opened by vtghsr 11 months ago - 2 comments

#579 - Error when training KeyError: "NuScenesDataset: 'infos'"

Issue - State: closed - Opened by EpicGilgamesh 11 months ago - 4 comments

#577 - Dimension inconsistency in convolusion operations in Conv2d and F.conv2d

Issue - State: closed - Opened by alexzhuuuu 11 months ago - 2 comments

#576 - Question about the Position of Lidar Backbone in the Model Architecture

Issue - State: closed - Opened by barrydoooit 11 months ago - 1 comment

#573 - changing point_cloud_range causes an error

Issue - State: closed - Opened by atto-js 11 months ago - 7 comments

#570 - Using a single image as input and limiting the field of view of the point cloud (nuScenes)

Issue - State: closed - Opened by J4nekT 11 months ago - 7 comments

#567 - there is a error

Issue - State: closed - Opened by winky4111 11 months ago - 1 comment

#560 - ModuleNotFoundError: No module named 'torchpack.utils.tqdm'

Issue - State: closed - Opened by Simeon340703 12 months ago - 2 comments

#559 - All types object class AP=0

Issue - State: closed - Opened by YHao29 12 months ago - 1 comment

#558 - How to train model with camera backbone ResNet50?

Issue - State: closed - Opened by FantineFu 12 months ago - 3 comments

#556 - error for test seg task

Issue - State: closed - Opened by Excutioner 12 months ago - 3 comments

#555 - visualize.py - savefig

Issue - State: closed - Opened by VeeranjaneyuluToka 12 months ago - 1 comment

#554 - Other pretrained models

Issue - State: closed - Opened by callenmacphee 12 months ago - 2 comments

#552 - nuscenes-mini Performance on TestSet Lower than the official results

Issue - State: closed - Opened by wstzx almost 1 year ago - 3 comments

#551 - Training with torchpack is unresponsive

Issue - State: closed - Opened by 1180100407 about 1 year ago - 3 comments

#550 - About Training Strategies

Issue - State: closed - Opened by lhcezx about 1 year ago - 1 comment

#545 - Question about `score_threshold` in centernet head config

Issue - State: closed - Opened by fligit about 1 year ago - 1 comment

#541 - Questions about Camera + Radar

Issue - State: closed - Opened by Terencedu about 1 year ago - 2 comments

#540 - ValueError: matrix contains invalid numeric entries

Issue - State: closed - Opened by 1653042420 about 1 year ago - 7 comments

#537 - question about depthss

Issue - State: closed - Opened by yu8ri1 about 1 year ago - 1 comment

#536 - Can I get your pre-trained model for R+C?

Issue - State: closed - Opened by geonhobang about 1 year ago - 1 comment

#534 - Tensor size mismatch

Issue - State: closed - Opened by AlexIlis about 1 year ago - 4 comments

#532 - TypeError: get_cam_feats() takes 3 positional arguments but 4 were given

Issue - State: closed - Opened by WangHeng1021 about 1 year ago - 14 comments

#531 - the MAP is very low,

Issue - State: closed - Opened by mzy991215 about 1 year ago - 3 comments

#527 - Primary job terminated normally, but 1 process returned

Issue - State: closed - Opened by ChirsHuanhuan about 1 year ago - 1 comment

#523 - how to fix the 'MMCV CUDA Compiler': 'not available' error?

Issue - State: closed - Opened by wangjunquan911 about 1 year ago - 2 comments

#522 - val map=0 error?

Issue - State: closed - Opened by mzy991215 about 1 year ago - 3 comments

#520 - why downsample factor assertion?

Issue - State: closed - Opened by zhaowenZhou about 1 year ago - 1 comment

#517 - KeyError: 'radar'`

Issue - State: closed - Opened by chg0901 about 1 year ago - 4 comments

#514 - Location key not present in main branch

Issue - State: closed - Opened by CSautier about 1 year ago - 2 comments

#512 - I am running the official training and have encountered the following error, which has caused me to be unable to train properly. Can anyone help me?

Issue - State: closed - Opened by lymcs about 1 year ago - 2 comments

#509 - [ROCm] add support for ROCm/HIP

Pull Request - State: closed - Opened by jeffdaily about 1 year ago

#506 - SECOND format

Issue - State: closed - Opened by Jayden9912 about 1 year ago - 2 comments

#505 - Nuscenes detection errors on validation set

Issue - State: closed - Opened by donand about 1 year ago - 1 comment

#504 - ImportError: cannot import name 'feature_decorator_ext' from partially initialized module 'mmdet3d.ops.feature_decorator' (most likely due to a circular import) (/data/bevfusion/mmdet3d/ops/feature_decorator/init.py)

Issue - State: closed - Opened by wangjunquan911 about 1 year ago - 7 comments

#503 - compilation terminated. error: command '/usr/bin/gcc' failed with exit code 1

Issue - State: closed - Opened by nican2018 about 1 year ago - 6 comments

#501 - When using single image (e.g. KITTI)

Issue - State: closed - Opened by san9569 about 1 year ago - 33 comments

#500 - TypeError: NuScenesDataset: init() missing 2 required positional arguments: 'doc' and 'pos'

Issue - State: closed - Opened by huzhihen about 1 year ago - 1 comment

#498 - I want to know the meaning of 'sensor2lidar', 'camera2lidar' parameter

Issue - State: closed - Opened by SeungHwi0613 about 1 year ago - 1 comment

#497 - Evalution RuntimeError: cusolver error: CUSOLVER_STATUS_INTERNAL_ERROR, when calling `cusolverDnCreate(handle)`

Issue - State: closed - Opened by cloudlockheart about 1 year ago - 6 comments

#489 - Update the instruction of the -np flag.

Pull Request - State: closed - Opened by DKandrew about 1 year ago

#488 - TypeError: 'DataContainer' object is not iterable

Issue - State: closed - Opened by Anna1ximi over 1 year ago - 4 comments

#486 - How much GPU memory is needed to train BEVFusion's model

Issue - State: closed - Opened by MaxLeo-LMZ over 1 year ago - 8 comments

#484 - can't use np>1

Issue - State: closed - Opened by Anna1ximi over 1 year ago - 8 comments

#483 - 'dict' object has no attribute 'rotate'

Issue - State: closed - Opened by quantumdotsss over 1 year ago - 3 comments

#478 - 7. RuntimeError: Given groups=1, weight of size [8, 1, 1, 1], expected input[12, 6, 256, 704] to have 1 channels, but got 6 channels instead

Issue - State: closed - Opened by MaxLeo-LMZ over 1 year ago - 20 comments

#471 - mAP=0 when testing on Night data (per class mAP)

Issue - State: closed - Opened by MinkyoungCho over 1 year ago - 1 comment

#470 - Warning: could not find environment variable "-x"

Issue - State: closed - Opened by wyf0414 over 1 year ago - 6 comments

#463 - The training results are relatively low.

Issue - State: closed - Opened by huzhihen over 1 year ago - 20 comments

#454 - updated devcontainer.json to support the gpu usage in codespace

Pull Request - State: closed - Opened by Unicorncosmos over 1 year ago

#445 - Questions about variants of models in the nuScenes test

Issue - State: closed - Opened by zoomkinseok over 1 year ago - 1 comment

#442 - visualizing ground truth

Issue - State: closed - Opened by VeeranjaneyuluToka over 1 year ago - 2 comments

GitHub / mit-han-lab/bevfusion issues and pull requests