yzd-v/FGD

Issues

• 你好，我看到您在GitHub更新了yolox的相关蒸馏的数据

  #94 opened 3 months ago by Northcandle
  4

• 你好，有空可以更新一下代码到mmdetection3.0以上版本吗。谢谢了

  #98 opened 3 months ago by chenjianhua1997
  0

• fgd.py文件中，area的值为0

  #97 opened 4 months ago by TinyOD
  0

• hello，I want to ask a question about yolox distillation

  #85 opened 2 years ago by yingwenmingzi123
  8

• To know Pytorch and cuda version

  #96 opened 5 months ago by nandini211995
  0

• FGD for Ultralytics frameworks

  #95 opened 6 months ago by DeeperNCheaper
  0

• Some questions about gcblock

  #93 opened a year ago by minhhotboy9x
  1

• Whether additional monitoring training is required for gc-block？

  #79 opened 2 years ago by onepeachbiubiubiu
  2

• How to output the Params or FLOPs after FGD distillation?

  #92 opened a year ago by chenqqqqqqq
  5

• Drop in mAP after Retinanet Distillation

  #91 opened a year ago by Rrschch-6
  10

• Can the network be replicated using the 2023 version of mmdetection?

  #89 opened a year ago by kiven228
  4

• YoloX-l distill Yolox-s with coco pretrained weights

  #90 opened a year ago by xiaobiaodu
  1

• Why the mAP is lower after distillation in VOC?

  #88 opened a year ago by cmjkqyl
  1

• invalid pre-training

  #87 opened a year ago by cmjkqyl
  7

• KeyError: 'neck.fpn_convs.4.conv'

  #86 opened 2 years ago by 0ppugno
  0

• mAP为0

  #71 opened 2 years ago by baibaidedongdong
  17

• The cfg distillation for yoloxl_m

  #74 opened 2 years ago by ALPAHz
  2

• Welcome update to OpenMMLab 2.0

  #84 opened 2 years ago by vansin
  0

• 请问作者有在yolov5上实现效果吗

  #83 opened 2 years ago by yeweliang
  2

• ConfigDict has no attribute "model"

  #82 opened 2 years ago by truong11062002
  6

• 1

  #81 opened 2 years ago by TomisJerry
  1

• The pre-trained model for Solo_r101_3x.pth

  #80 opened 2 years ago by sscckk
  1

• no module named 'mmdet.distillation'

  #78 opened 2 years ago by BonBon-Boom
  1

• How Can I install mmdet 2.11.0 while its latest version is 2.28.1?

  #77 opened 2 years ago by BonBon-Boom
  1

• Using fgd_retina_r101_fpn_2x_distill_retina_r50_fpn_2x_coco.py to find retina_ R50 distillation result AP is 37.3%?

  #75 opened 2 years ago by wangtao2668129173
  1

• Why does inference slow down with weights generated by FGD？

  #76 opened 2 years ago by cmjkqyl
  1

• Train my self data, the loss is large

  #65 opened 2 years ago by baibaidedongdong
  6

• Add DCN

  #73 opened 2 years ago by ywaner
  5

• Cascadercnn-resnet50蒸馏18,loss_fgd_fpn_0太大（第一张图）

  #72 opened 2 years ago by Bingchuan521
  10

• 为什么我运行的时候总提示 the model and loaded state dict do not match exactly?

  #70 opened 2 years ago by chenqq1
  2

• The model and loaded state dict do not match exactly

  #69 opened 2 years ago by chenqq1
  2

• Why the mAP is 0 when train myself data?

  #68 opened 2 years ago by baibaidedongdong
  1

• FCOS训练及测试结果问题

  #66 opened 2 years ago by Hwang64
  4

• Should I change the mmdet/detectors/base.py when I train yolox?

  #67 opened 2 years ago by baibaidedongdong
  1

• the

  #64 opened 2 years ago by baibaidedongdong
  0

• if i can use MultiScale training？

  #63 opened 2 years ago by baibaidedongdong
  3

• Warnings during the training process for FCOS

  #62 opened 2 years ago by Hwang64
  1

• what does it means

  #61 opened 2 years ago by baibaidedongdong
  1

• FCOS实验结果问题

  #60 opened 2 years ago by changxin111
  5

• Using swin model as teacher

  #59 opened 2 years ago by peiyingxin
  0

• 实验数据是否有误？感谢

  #58 opened 2 years ago by RY-97
  2

• Hyper-parameter settings on ablation study

  #56 opened 2 years ago by jwfanDL
  1

• A question about object detection distill

  #57 opened 2 years ago by peiyingxin
  0

• How to load a yolox-l correctly?

  #54 opened 2 years ago by Errol-golang
  1

• 利用 FGD将两个教师模型训练一个学生模型的问题

  #55 opened 2 years ago by lixiangMindSpore
  0

• About the cfg distillation for yolox

  #53 opened 2 years ago by Errol-golang
  5

• Testing result

  #52 opened 2 years ago by peiyingxin
  6

• fgd_yoloxl_distill_yoloxm_coco.py中的FGDLoss就是fgd_solo_r101_ms_distill_solo_r50_coco.py中的FeatureLoss吗？

  #51 opened 2 years ago by lixiangMindSpore
  2

• About result

  #50 opened 2 years ago by peiyingxin
  1

• No module named 'mmdet.distillation'

  #49 opened 2 years ago by lixiangMindSpore
  3