YuwenXiong/py-R-FCN

Issues

• coco training problem

  #85 opened 7 years ago by 2017hack
  11

• Message type "caffe.LayerParameter" has no field named "psroi_pooling_param".

  #128 opened 4 years ago by wanghangege
  0

• question

  #127 opened 5 years ago by 13668379467
  0

• Could you give me some hints to fix this issue? Thanks a lot!

  #126 opened 5 years ago by newforrestgump001
  0

• How to deal with gt_boxes zero sample?>

  #125 opened 5 years ago by lucasjinreal
  0

• How to set iterations when training with rfcn_alt_opt_5stage_ohem

  #124 opened 6 years ago by wander1985
  0

• About depth of RoI-wise subnetwork

  #123 opened 6 years ago by sjtuxcx
  0

• undefined reference to symbol 'cudaSetDevice'

  #122 opened 6 years ago by ihg1992
  0

• It is goning wrong when i load the pre_train model!!!

  #104 opened 7 years ago by Zzmc
  1

• train ResNet-50 Alternating training Erroe

  #119 opened 6 years ago by chenxu93
  3

• eltwise_layer.cpp:34 check failed: bottom[i]->shape() == bottom[0]->shape(). error

  #121 opened 6 years ago by daiwenjun2017
  0

• when i train rfcn_end2end,there is an error in proposal_target_layer.py

  #120 opened 6 years ago by zhaowujie
  0

• Implement Light-Head RCNN by modifying the src of RFCN

  #101 opened 7 years ago by sherryxie1
  2

• a implementation of light head rcnn by modifying R-FCN, low mAP

  #118 opened 6 years ago by ZhiHuaYin
  0

• How to implement R-FCN using ZF or VGG base model

  #100 opened 7 years ago by liuliu66
  1

• Issue with building microsoft caffe

  #117 opened 6 years ago by liminghuiv
  0

• Why BN/Scale layers are not included in fine-tune?

  #116 opened 7 years ago by ltqusst
  0

• Is the backbone network never fine-tuned during training?

  #115 opened 7 years ago by ltqusst
  1

• ValueError: total size of new array must be unchanged

  #114 opened 7 years ago by clxia12
  0

• about multi-scale traing/testing

  #113 opened 7 years ago by foralliance
  0

• about lib/fast_rcnn/train.py

  #112 opened 7 years ago by foralliance
  0

• Why is cfg.TRAIN.USE_OHEM=False for OHEM training?

  #103 opened 7 years ago by kampelmuehler
  1

• why the training was interrupted unexpectedly after some iteration(60000)?

  #102 opened 7 years ago by shadowuyl
  1

• how to generate a solverstate?

  #95 opened 7 years ago by muyue1238
  2

• I changed it to a single category detection and the loss become NONE in train time after 1000 iter

  #111 opened 7 years ago by Chen94yue
  0

• error use caffe-Microsoft

  #110 opened 7 years ago by wangjy2020
  0

• Very low test accuracy using OHEM.

  #91 opened 7 years ago by MissDores
  3

• how to solve the lib reimplement for minibatch size >1 error?

  #109 opened 7 years ago by ygren
  0

• what's the difference between class-aware and agnostic?

  #107 opened 7 years ago by scut-salmon
  6

• bbox_loss is always zero when I train R-FCN with OHEM

  #108 opened 7 years ago by starxhong
  1

• what's the mean of "alternative optimization" for the result table

  #106 opened 7 years ago by scut-salmon
  1

• How to train the model with class aware box?

  #105 opened 7 years ago by panovr
  0

• how to using python api to generate your https://github.com/YuwenXiong/py-R-FCN/blob/master/models/pascal_voc/ResNet-50/rfcn_end2end/train_agnostic.prototxt

  #99 opened 7 years ago by qinyi20060410
  1

• Can't parse message of type "caffe.NetParameter"

  #97 opened 7 years ago by lucasjinreal
  1

• train error

  #93 opened 7 years ago by wjwstate
  1

• SmoothL1LossOHEMParameter this layer not in caffe.proto

  #98 opened 7 years ago by lucasjinreal
  1

• Why is the output_dim of psroipooled_loc_rois is 8? Why not 4?

  #96 opened 7 years ago by MissLikeWind
  0

• Merged folder for image set?

  #94 opened 7 years ago by chriszhenghaochen
  2

• Is the data still available ?

  #89 opened 7 years ago by bibinjose
  0

• extension to arbitrary number of channel

  #92 opened 7 years ago by laupl0082
  0

• What's the version of CUDNN should be used?

  #90 opened 7 years ago by MissDores
  1

• Problem: Training using 5 stage mode

  #88 opened 7 years ago by xxxxxxxx-dl
  0

• joint training vers alternative optimization

  #86 opened 7 years ago by 2017hack
  1

• Multi-GPU Support

  #76 opened 7 years ago by yehan-xiao
  1

• Why the output of "rfcn_bbox" layer is 392?

  #83 opened 7 years ago by panovr
  1

• Ignore Difficult_Case set to True

  #82 opened 7 years ago by ananddb90
  1

• Is the alternative-optimization result is trained with OHEM?

  #81 opened 7 years ago by Zehaos
  3

• peformance on smaller network

  #78 opened 8 years ago by newwhitecheng
  3

• Using the class-agnostic. why the bbox num_output also need cls_num*4?

  #80 opened 7 years ago by 315386775
  1

• Whether I can use this code on windows10?

  #79 opened 7 years ago by xdsonglinliu
  2