chuanqi305/MobileNet-SSD

Issues

• add feature map 38x38

  #98 opened a year ago by OMG59E
  2

• A small advice: update the readme

  #118 opened 3 years ago by zchrissirhcz
  0

• 检测结果：置信度和类别为-1的情况

  #137 opened 6 years ago by lvpchen
  2

• 传奇你好，我将ssd结构改成sfd并使用wider face数据集做人脸检测的时候，一开始loss是收敛的（刚开始大概），然后在loss=8的时候震荡

  #100 opened 7 years ago by wanghanyang123
  41

• error in running merge_bn.py

  #131 opened 6 years ago by meixiansen
  5

• How to resume training?

  #127 opened 6 years ago by tolotrasamuel
  1

• why don't convolution depth wise use engine: CUDNN instead of engine: Caffe?

  #122 opened 6 years ago by ZhiqiJiang
  1

• Why MobileNet-SSD don't have 'step' this variable?

  #119 opened 6 years ago by JingliangGao
  1

• how to run it with high speed in tx2

  #148 opened 6 years ago by Eutenacity
  1

• MobileNet-SSD: Assertion `min_arg < max_arg' failed.

  #112 opened 6 years ago by Dudule33
  5

• Is it possible to replace image preprocess(normalization) with a 1x1 conv layer?

  #136 opened 6 years ago by RunshengZhu
  1

• Error when i use merge.py

  #134 opened 6 years ago by pluie
  2

• train error "Error Incompatible number of blobs for layer conv0"

  #145 opened 6 years ago by JonnySme
  1

• Iteration Vs Number of images Vs Batch size and Accuracy?

  #141 opened 6 years ago by PiyalGeorge
  0

• How to reduce delay in Real-time Face detection?

  #109 opened 6 years ago by PiyalGeorge
  0

• How to do Face Recognition?

  #108 opened 6 years ago by PiyalGeorge
  0

• Has anyone tried mobilenet-ssd on TX2,what`s the performance?

  #132 opened 6 years ago by zbw4034
  3

• opencv error

  #105 opened 7 years ago by cherryxiongyw
  3

• [data_transformer.cpp:1036] Check failed: datum_channels > 0 (0 vs. 0)

  #147 opened 6 years ago by Imraj
  1

• Check failed: datum_channels > 0

  #110 opened 6 years ago by hsweif
  1

• std::length_error on train

  #146 opened 6 years ago by FSet89
  0

• transform dada shape is zero

  #144 opened 6 years ago by zhoupan9109
  0

• How can I use the single channel image as the input data?

  #143 opened 6 years ago by zhoupan9109
  0

• training model with no bias very slow to converge?

  #142 opened 6 years ago by ndcuong91
  0

• The loss function converges slowly and the test accuracy is not high??

  #140 opened 6 years ago by lunalulu
  0

• How can i modify the input size?

  #139 opened 6 years ago by pluie
  0

• MobileNet-SSD inputsize192x192

  #99 opened 7 years ago by caoqichun
  2

• merge_bn.py运行出现protobuf问题

  #138 opened 6 years ago by nawei87
  0

• Merge_bn accuracy 0

  #135 opened 6 years ago by JinyangGuo
  0

• add conversion script from tensorflow to caffemodel?

  #133 opened 6 years ago by kodonnell
  2

• false detection as negative samples

  #130 opened 6 years ago by lvpchen
  0

• Cannot find lmdb when training

  #124 opened 6 years ago by afantideng
  1

• forward time and model size

  #128 opened 6 years ago by yja1
  1

• Killed when run on jetson TX2

  #126 opened 6 years ago by FelixCaae
  2

• Does test.sh evaluate mAP of detection?

  #111 opened 6 years ago by voqtuyen
  1

• How many flops does this use?

  #125 opened 6 years ago by hazemati
  0

• How to reduce the number of classes ?

  #117 opened 6 years ago by tolotrasamuel
  1

• training stuck at iter 0 and stops without throwing any error ,i used docker image cpu on google cloud,

  #123 opened 6 years ago by manoj652
  0

• failed to upgrade deploy.prototxt while using caffe's upgrade_net_proto_text tool

  #121 opened 6 years ago by xiaofengren
  0

• Try training the MobileNet SSD network for Caltech Dataset, but keeps got loss = 0.

  #116 opened 6 years ago by DanielXu123
  2

• Using the compute_image_mean script getting 8 bytes binaryproto file

  #120 opened 6 years ago by DanielXu123
  0

• Good news: a window version of caffe supporting MobileNet-SSD

  #115 opened 6 years ago by zuoqing1988
  0

• Where can we find the model trained with MSCOCO dataset only ? (90 classes)

  #114 opened 6 years ago by tolotrasamuel
  0

• Why after the iteration of 20000 times, the loss oscillates around 1.4, no longer reduces

  #113 opened 6 years ago by HappyKerry
  1

• run demo.py, the result was incorrect

  #107 opened 6 years ago by lunalulu
  1

• can `gen.py` generate `deploy` without `bn` prototxt?

  #106 opened 6 years ago by zimenglan-sysu-512
  0

• caffe.NetParameter: 1177:17: Message type "caffe.LayerParameter" has no field named "permute_param".

  #102 opened 7 years ago by s0r2637
  1

• 256x256 input size and question about pre-processing of images

  #104 opened 7 years ago by Cathynuo
  0

• how to finetune MobileNet-SSD?

  #103 opened 7 years ago by xufeifeiWHU
  0

• Accuracy Testing

  #101 opened 7 years ago by arisha07
  0