Sense-X/TSD

Issues

• How to implement Naive Ensemble and Auto Ensemble

  #40 opened 3 months ago by ckx506772099
  0

• TypeError "forward() missing 2 required positional arguments: 'feats' and 'rois'" in pytorch2onnx

  #34 opened 4 years ago by esdream
  1

• Pretrained weights for OpenImages dataset (600 classes)?

  #14 opened 4 years ago by kiranramnath007
  9

• Welcome update to OpenMMLab 2.0

  #39 opened 2 years ago by vansin
  0

• TypeError: FasterRCNN: __init__() got an unexpected keyword argument 'bbox_roi_extractor'

  #38 opened 3 years ago by mosheliv
  1

• Question about OpenImages evaluation

  #35 opened 3 years ago by BIGWangYuDong
  0

• How to use it with A100 Gpus?

  #37 opened 3 years ago by HenryHZY
  0

• Could you share the training log and evaluation script？

  #36 opened 3 years ago by howardchina
  0

• RuntimeError: cuda runtime error (11) : invalid argument at mmdet/ops/nms/src/nms_kernel.cu:111

  #32 opened 3 years ago by hq03
  0

• How to illustrate the heatmap thermal diagram?

  #27 opened 3 years ago by Derek-Kun
  1

• Spatial sensitive heatmap

  #25 opened 4 years ago by DingShengLin
  4

• Open-Images datasets

  #33 opened 4 years ago by cailk
  1

• I transfer the code to mmdetection v2.0, but the result is much worse than general training pipeline. What might be the reason？

  #31 opened 4 years ago by clw5180
  3

• Missing 'model.CLASSES' field in the pre-trained model for the open images dataset

  #30 opened 4 years ago by iMayK
  3

• KeyError: 'TSDSharedFCBBoxHead is not in the head registry'

  #29 opened 4 years ago by iMayK
  1

• Can share the codes about Sampling and ensembling?

  #1 opened 4 years ago by shipengai
  5

• Some questions about TSD_cls and TSD_bbox

  #20 opened 4 years ago by wangjue-wzq
  9

• Can TSD head be used on top of RetinaNet?

  #19 opened 4 years ago by CodeJjang
  5

• DDP problem when migrate TSD to mmdet2.0

  #26 opened 4 years ago by zhaoxin111
  0

• It seems that the mmcv version does not match.

  #17 opened 4 years ago by limengfei3675
  5

• The x101_64x4d model of train and inference is very very slow?

  #6 opened 5 years ago by lexiqi
  3

• Should not overwrite two_stage.py

  #24 opened 4 years ago by ylf11235
  1

• Why the rcnn score_thr is set 0.00 in all TSD configs?

  #23 opened 4 years ago by tmp12316
  0

• mmcv version

  #22 opened 4 years ago by yangshudi
  2

• Could you share the time cost of train and test, comparing to without TSD?

  #13 opened 4 years ago by lint39
  1

• Question about how to set finest_scale value?

  #21 opened 4 years ago by wangjue-wzq
  2

• About the confidence threshold

  #18 opened 4 years ago by wangerxiao001
  2

• inconsistency of mmcv or mmdetection?

  #16 opened 4 years ago by chengjiun
  1

• InstallationError: nvcc fatal : Unsupported gpu architecture 'compute_75' ninja: build stopped: subcommand failed. Error compiling objects for extension

  #15 opened 4 years ago by raoufkeskes
  1

• UnboundLocalError: local variable 'pc_cls_loss' referenced before assignment

  #12 opened 4 years ago by fengepsilon
  2

• Can TSDSharedFCBBoxHead work fine in Cascade?

  #7 opened 4 years ago by luuuyi
  7

• i can train your model(ap 49.2)successfully on my own datasets,but it doesn't output eval info per epoch

  #9 opened 4 years ago by kiddliao
  3

• Can this method be used for feature alignment of one-stage object detector? Such as SSD and YOLO?

  #8 opened 4 years ago by feixiangdekaka
  1

• tsd

  #10 opened 4 years ago by lijain
  1

• Detailed evaluation results on OID

  #11 opened 4 years ago by Jokoe66
  1

• rcnn nms score_thre？

  #5 opened 5 years ago by luuuyi
  2

• An error occurred while using fp16

  #2 opened 5 years ago by Dragonsson
  3

• Question about training with fewer GPUs

  #4 opened 5 years ago by Scalsol
  2

• loss function

  #3 opened 5 years ago by dereyly
  1