search

corleonechensiyu / AdvancedEAST-caffe

caffe 实现 AdvancedEAST+mobilenetv3
4 stars 3 forks source link

mbv3\train.prototxt #1

Open polooloo opened 4 years ago

polooloo commented 4 years ago

This mbv3 / train.prototxt cannot be used. It should be in this format under normal circumstances: Should include TRAIN and TEST parsing parameters. Please correct it as soon as possible.

layer {    name: "data"    type: "Data"    top: "data"    top: "label"    include {      phase: TRAIN    }   transform_param {     mean_value: 132     mean_value: 128     mean_value: 128     mirror: true   }    data_param {      source: "D: / BLTDeep / MobileNetV3 / caffe-windows / models / EfficientNet-Caffe / lmdb / train"      batch_size: 8      backend: LMDB    } } layer {    name: "data"    type: "Data"    top: "data"    top: "label"    include {      phase: TEST    }

mean pixel / channel-wise mean instead of mean image

  transform_param   {     mean_value: 132     mean_value: 128     mean_value: 128     mirror: false   }    data_param {      source: "D: / BLTDeep / MobileNetV3 / caffe-windows / models / EfficientNet-Caffe / lmdb / test"      batch_size: 8      backend: LMDB    } }

polooloo commented 4 years ago

I tested it according to your directory structure and found that I can't train the network. Your train.prototxt should contain at least the input and output data sets that need to be trained, right?

corleonechensiyu commented 4 years ago

I tested it according to your directory structure and found that I can't train the network. Your train.prototxt should contain at least the input and output data sets that need to be trained, right?

训练文件不全，loss函数领有文件，这几天在准备毕业论文的事，我会全部公开，包括caffe怎样用Python写自己的loss函数等一些经验教程

polooloo commented 4 years ago

Could you please provide me with a full training profile?

polooloo commented 4 years ago

I tested it according to your directory structure and found that I can't train the network. Your train.prototxt should contain at least the input and output data sets that need to be trained, right?

训练文件不全，loss函数领有文件，这几天在准备毕业论文的事，我会全部公开，包括caffe怎样用Python写自己的loss函数等一些经验教程

Could you please provide me with a full training profile?

corleonechensiyu commented 4 years ago

I tested it according to your directory structure and found that I can't train the network. Your train.prototxt should contain at least the input and output data sets that need to be trained, right?

训练文件不全，loss函数领有文件，这几天在准备毕业论文的事，我会全部公开，包括caffe怎样用Python写自己的loss函数等一些经验教程

Could you please provide me with a full training profile?

In a few days

polooloo commented 4 years ago

I tested it according to your directory structure and found that I can't train the network. Your train.prototxt should contain at least the input and output data sets that need to be trained, right?

训练文件不全，loss函数领有文件，这几天在准备毕业论文的事，我会全部公开，包括caffe怎样用Python写自己的loss函数等一些经验教程

Could you please provide me with a full training profile?

In a few days

Looking forward to the guidance of the Great God!