hito0512 commented 7 years ago

i have change anchor ,and change num_output as follow: layer { name: "rpn_cls_score" type: "Convolution" bottom: "rpn/output" top: "rpn_cls_score" param { lr_mult: 1.0 } param { lr_mult: 2.0 } convolution_param {

num_output: 18 # 2(bg/fg) * 9(anchors)

num_output: 40 # 2(bg/fg) * 20(anchors) kernel_size: 1 pad: 0 stride: 1 weight_filler { type: "gaussian" std: 0.01 } bias_filler { type: "constant" value: 0 } } } layer { name: "rpn_bbox_pred" type: "Convolution" bottom: "rpn/output" top: "rpn_bbox_pred" param { lr_mult: 1.0 } param { lr_mult: 2.0 } convolution_param {

num_output: 36 # 4 * 9(anchors)

num_output: 80   # 4 * 20(anchors)
kernel_size: 1 pad: 0 stride: 1
weight_filler { type: "gaussian" std: 0.01 }
bias_filler { type: "constant" value: 0 }

}`

layer { name: 'rpn_cls_prob_reshape' type: 'Reshape' bottom: 'rpn_cls_prob' top: 'rpn_cls_prob_reshape' reshape_param { shape { dim: 0 dim: 40 dim: -1 dim: 0 } }
}

but when i run demo the terminal also show layer { name: "rpn_cls_score" type: "Convolution" bottom: "rpn_conv1" top: "rpn_cls_score" convolution_param { num_output: 18 pad: 0 kernel_size: 1 stride: 1 } } layer { name: "rpn_bbox_pred" type: "Convolution" bottom: "rpn_conv1" top: "rpn_bbox_pred" convolution_param { num_output: 36 pad: 0 kernel_size: 1 stride: 1 } } F0603 20:02:36.350150 5923 net.cpp:829] Cannot copy param 0 weights from layer 'rpn_cls_score'; shape mismatch. Source param shape is 40 256 1 1 (10240); target param shape is 18 256 1 1 (4608). To learn this layer's parameters from scratch rather than copying from a saved net, rename the layer. Check failure stack trace: why????

,please help me .

djdam commented 7 years ago

you're loading the weights from a network that had 18 anchors instead of 40, so it cannot match the amount of weights with those of your network. Like the error says, rename your layer (e.g. rpn_cls_score2) and then run again. This layer is then trained from scratch.

hito0512 commented 7 years ago

@djdam thx ,that problem have solved,but i rename layer name and change anchor (base_size=16, ratios=[0.3,0.5,1],scales=[2,4,8,16]): and anchor_scales = layer_params.get('scales', (2,4,8,16))

layer { name: "rpn_cls_score_0" type: "Convolution" bottom: "rpn/output" top: "rpn_cls_score" convolution_param { num_output: 24

layer { name: "rpn_bbox_pred_0" type: "Convolution" bottom: "rpn/output" top: "rpn_bbox_pred" convolution_param { num_output: 48

layer { name: 'rpn_cls_prob_reshape_0' type: 'Reshape' bottom: 'rpn_cls_prob' top: 'rpn_cls_prob_reshape' reshape_param { shape { dim: 0 dim: 24 dim: -1 dim: 0 } } #原来 18 } i get other error 2017-06-04 17-06-16 i dont know why ,i dont know what should i do

djdam commented 7 years ago

Hi Hito

try to delete the cached .pkl file, usually located at py-faster-rcnn/data/cache/train__gt_roidb.pkl

hito0512 commented 7 years ago

@djdam thx,it is the proposal_layer.py, if you change anchor_scales = layer_params.get('scales',(2,4,8,16)) like this ,it occur mistake,but if you change anchor_scales = layer_params.get('scales',(2,4,8)) like this ,it is right , anchor_scales only contain 3 parameters, i dont know why

djdam commented 7 years ago

@jiayandekafei can you post the complete prototxt?

hito0512 commented 7 years ago

@djdam train.prototxt

name: "ZF" layer { name: 'input-data' type: 'Python' top: 'data' top: 'im_info' top: 'gt_boxes' python_param { module: 'roi_data_layer.layer' layer: 'RoIDataLayer' param_str: "'num_classes': 4" #按训练集类别改，该值为类别数+1 } }

========= conv1-conv5 ============

layer { name: "conv1" type: "Convolution" bottom: "data" top: "conv1" param { lr_mult: 1.0 } param { lr_mult: 2.0 } convolution_param { num_output: 96 kernel_size: 7 pad: 3 stride: 2 } } layer { name: "relu1" type: "ReLU" bottom: "conv1" top: "conv1" } layer { name: "norm1" type: "LRN" bottom: "conv1" top: "norm1" lrn_param { local_size: 3 alpha: 0.00005 beta: 0.75 norm_region: WITHIN_CHANNEL engine: CAFFE } } layer { name: "pool1" type: "Pooling" bottom: "norm1" top: "pool1" pooling_param { kernel_size: 3 stride: 2 pad: 1 pool: MAX } } layer { name: "conv2" type: "Convolution" bottom: "pool1" top: "conv2" param { lr_mult: 1.0 } param { lr_mult: 2.0 } convolution_param { num_output: 256 kernel_size: 5 pad: 2 stride: 2 } } layer {
name: "relu2" type: "ReLU" bottom: "conv2" top: "conv2" } layer { name: "norm2" type: "LRN" bottom: "conv2" top: "norm2" lrn_param { local_size: 3 alpha: 0.00005 beta: 0.75 norm_region: WITHIN_CHANNEL engine: CAFFE } }

================ #将此池化层删除 ===============

layer { name: "pool2"
type: "Pooling" bottom: "norm2" top: "pool2" pooling_param { kernel_size: 3 stride: 2 pad: 1 pool: MAX } }

===========================================

layer { name: "conv3" type: "Convolution" bottom: "pool2" # 原来输入为pool2 top: "conv3" param { lr_mult: 1.0 } param { lr_mult: 2.0 } convolution_param { num_output: 384 kernel_size: 3 pad: 1 stride: 1 } } layer { name: "relu3" type: "ReLU" bottom: "conv3" top: "conv3" } layer { name: "conv4" type: "Convolution" bottom: "conv3" #conv3 top: "conv4" param { lr_mult: 1.0 } param { lr_mult: 2.0 } convolution_param { num_output: 384 kernel_size: 3 pad: 1 stride: 1 } } layer { name: "relu4" type: "ReLU" bottom: "conv4" top: "conv4" } layer { name: "conv5" type: "Convolution" bottom: "conv4" top: "conv5" param { lr_mult: 1.0 } param { lr_mult: 2.0 } convolution_param { num_output: 256 kernel_size: 3 pad: 1 stride: 1 } } layer { name: "relu5" type: "ReLU" bottom: "conv5" top: "conv5" }

========= RPN ============

layer { name: "rpn_conv/3x3" type: "Convolution" bottom: "conv5" top: "rpn/output" param { lr_mult: 1.0 } param { lr_mult: 2.0 } convolution_param { num_output: 256 kernel_size: 3 pad: 1 stride: 1 weight_filler { type: "gaussian" std: 0.01 } bias_filler { type: "constant" value: 0 } } } layer { name: "rpn_relu/3x3" type: "ReLU" bottom: "rpn/output" top: "rpn/output" }

layer { name: "rpn_cls_score" type: "Convolution" bottom: "rpn/output" top: "rpn_cls_score" param { lr_mult: 1.0 } param { lr_mult: 2.0 } convolution_param { num_output: 24 # 2(bg/fg) * 9(anchors)

num_output: 40 # 2(bg/fg) * 20(anchors)

kernel_size: 1 pad: 0 stride: 1
weight_filler { type: "gaussian" std: 0.01 }
bias_filler { type: "constant" value: 0 }

} } layer { name: "rpn_bbox_pred" type: "Convolution" bottom: "rpn/output" top: "rpn_bbox_pred" param { lr_mult: 1.0 } param { lr_mult: 2.0 } convolution_param { num_output: 48 # 4 * 9(anchors)

num_output: 80 # 4 * 20(anchors)

kernel_size: 1 pad: 0 stride: 1
weight_filler { type: "gaussian" std: 0.01 }
bias_filler { type: "constant" value: 0 }

} } layer { bottom: "rpn_cls_score" top: "rpn_cls_score_reshape" name: "rpn_cls_score_reshape" type: "Reshape" reshape_param { shape { dim: 0 dim: 2 dim: -1 dim: 0 } } } layer { name: 'rpn-data' type: 'Python' bottom: 'rpn_cls_score' bottom: 'gt_boxes' bottom: 'im_info' bottom: 'data' top: 'rpn_labels' top: 'rpn_bbox_targets' top: 'rpn_bbox_inside_weights' top: 'rpn_bbox_outside_weights' python_param { module: 'rpn.anchor_target_layer' layer: 'AnchorTargetLayer' param_str: "'feat_stride': 16" #16 } } layer { name: "rpn_loss_cls" type: "SoftmaxWithLoss" bottom: "rpn_cls_score_reshape" bottom: "rpn_labels" propagate_down: 1 propagate_down: 0 top: "rpn_cls_loss" loss_weight: 1 loss_param { ignore_label: -1 normalize: true } } layer { name: "rpn_loss_bbox" type: "SmoothL1Loss" bottom: "rpn_bbox_pred" bottom: "rpn_bbox_targets" bottom: 'rpn_bbox_inside_weights' bottom: 'rpn_bbox_outside_weights' top: "rpn_loss_bbox" loss_weight: 1 smooth_l1_loss_param { sigma: 3.0 } }

========= RoI Proposal ============

layer { name: "rpn_cls_prob" type: "Softmax" bottom: "rpn_cls_score_reshape" top: "rpn_cls_prob" } layer { name: 'rpn_cls_prob_reshape' type: 'Reshape' bottom: 'rpn_cls_prob' top: 'rpn_cls_prob_reshape' reshape_param { shape { dim: 0 dim: 24 dim: -1 dim: 0 } } # dim: 0 dim: 18 dim: -1 dim: 0 layer { name: 'proposal' type: 'Python' bottom: 'rpn_cls_prob_reshape' bottom: 'rpn_bbox_pred' bottom: 'im_info' top: 'rpn_rois' python_param { module: 'rpn.proposal_layer' layer: 'ProposalLayer' param_str: "'feat_stride': 16" # param_str: "'feat_stride': 16" } }

layer { name: 'roi-data' type: 'Python' bottom: 'rpn_rois' bottom: 'gt_boxes' top: 'rois' top: 'labels' top: 'bbox_targets' top: 'bbox_inside_weights' top: 'bbox_outside_weights' python_param { module: 'rpn.proposal_target_layer' layer: 'ProposalTargetLayer' param_str: "'num_classes': 4" #按训练集类别改，该值为类别数+1 } }

========= RCNN ============

layer { name: "roi_pool_conv5" type: "ROIPooling" bottom: "conv5" bottom: "rois" top: "roi_pool_conv5" roi_pooling_param { pooled_w: 6 pooled_h: 6 spatial_scale: 0.0625 # 1/16 } } layer { name: "fc6" type: "InnerProduct" bottom: "roi_pool_conv5" top: "fc6" param { lr_mult: 1.0 } param { lr_mult: 2.0 } inner_product_param { num_output: 4096 } } layer { name: "relu6" type: "ReLU" bottom: "fc6" top: "fc6" } layer { name: "drop6" type: "Dropout" bottom: "fc6" top: "fc6" dropout_param { dropout_ratio: 0.5 scale_train: false } } layer { name: "fc7" type: "InnerProduct" bottom: "fc6" top: "fc7" param { lr_mult: 1.0 } param { lr_mult: 2.0 } inner_product_param { num_output: 4096 } } layer { name: "relu7" type: "ReLU" bottom: "fc7" top: "fc7" } layer { name: "drop7" type: "Dropout" bottom: "fc7" top: "fc7" dropout_param { dropout_ratio: 0.5 scale_train: false } }

layer { name: "cls_score_my" type: "InnerProduct" bottom: "fc7" top: "cls_score" #"cls_score" param { lr_mult: 1.0 } param { lr_mult: 2.0 } inner_product_param { num_output: 4 #按训练集类别改，该值为类别数+1 weight_filler { type: "gaussian" std: 0.01 } bias_filler { type: "constant" value: 0 } } }

layer { name: "bbox_pred_my" #bbox_pred type: "InnerProduct" bottom: "fc7" top: "bbox_pred" #bbox_pred param { lr_mult: 1.0 } param { lr_mult: 2.0 } inner_product_param { num_output: 16 #按训练集类别改，该值为（类别数+1）*4 weight_filler { type: "gaussian" std: 0.001 } bias_filler { type: "constant" value: 0 } } }

layer { name: "loss_cls" type: "SoftmaxWithLoss" bottom: "cls_score" #bottom: "cls_score" bottom: "labels" propagate_down: 1 propagate_down: 0 top: "cls_loss" loss_weight: 1 loss_param { ignore_label: -1 normalize: true } } layer { name: "loss_bbox" type: "SmoothL1Loss" bottom: "bbox_pred" bottom: "bbox_targets" bottom: 'bbox_inside_weights' bottom: 'bbox_outside_weights' top: "bbox_loss" loss_weight: 0.85 # }

test.prototxt name: "ZF"

input: "data" input_shape { dim: 1 dim: 3 dim: 224 dim: 224 }

input: "im_info" input_shape { dim: 1 dim: 3 }

========= conv1-conv5 ============

layer { name: "conv1" type: "Convolution" bottom: "data" top: "conv1" convolution_param { num_output: 96 kernel_size: 7 pad: 3 stride: 2 } } layer { name: "relu1" type: "ReLU" bottom: "conv1" top: "conv1" } layer { name: "norm1" type: "LRN" bottom: "conv1" top: "norm1" lrn_param { local_size: 3 alpha: 0.00005 beta: 0.75 norm_region: WITHIN_CHANNEL engine: CAFFE } } layer { name: "pool1" type: "Pooling" bottom: "norm1" top: "pool1" pooling_param { kernel_size: 3 stride: 2 pad: 1 pool: MAX } } layer { name: "conv2" type: "Convolution" bottom: "pool1" top: "conv2" convolution_param { num_output: 256 kernel_size: 5 pad: 2 stride: 2 } } layer { name: "relu2" type: "ReLU" bottom: "conv2" top: "conv2" } layer { name: "norm2" type: "LRN" bottom: "conv2" top: "norm2" lrn_param { local_size: 3 alpha: 0.00005 beta: 0.75 norm_region: WITHIN_CHANNEL engine: CAFFE } }

=========================================

layer { name: "pool2" type: "Pooling" bottom: "norm2" top: "pool2" pooling_param { kernel_size: 3 stride: 2 pad: 1 pool: MAX } }

=========================================

layer { name: "conv3" type: "Convolution" bottom: "pool2" #"pool2" top: "conv3" convolution_param { num_output: 384 kernel_size: 3 pad: 1 stride: 1 } } layer { name: "relu3" type: "ReLU" bottom: "conv3" top: "conv3" } layer { name: "conv4" type: "Convolution" bottom: "conv3" top: "conv4" convolution_param { num_output: 384 kernel_size: 3 pad: 1 stride: 1 } } layer { name: "relu4" type: "ReLU" bottom: "conv4" top: "conv4" } layer { name: "conv5" type: "Convolution" bottom: "conv4" top: "conv5" convolution_param { num_output: 256 kernel_size: 3 pad: 1 stride: 1 } } layer { name: "relu5" type: "ReLU" bottom: "conv5" top: "conv5" }

========= RPN ============

layer { name: "rpn_conv/3x3" type: "Convolution" bottom: "conv5" top: "rpn/output" convolution_param { num_output: 256 kernel_size: 3 pad: 1 stride: 1 weight_filler { type: "gaussian" std: 0.01 } bias_filler { type: "constant" value: 0 } } } layer { name: "rpn_relu/3x3" type: "ReLU" bottom: "rpn/output" top: "rpn/output" }

====================================

layer { name: "rpn_cls_score" type: "Convolution" bottom: "rpn/output" top: "rpn_cls_score" convolution_param { num_output: 24 # 2(bg/fg) * 9(anchors) #原来

num_output: 40 # 2(bg/fg) * 20(anchors)

kernel_size: 1 pad: 0 stride: 1
weight_filler { type: "gaussian" std: 0.01 }
bias_filler { type: "constant" value: 0 }

} } layer { name: "rpn_bbox_pred" type: "Convolution" bottom: "rpn/output" top: "rpn_bbox_pred" convolution_param { num_output: 48 # 4 * 9(anchors) #原来

num_output: 80 # 4 * 20(anchors)

kernel_size: 1 pad: 0 stride: 1
weight_filler { type: "gaussian" std: 0.01 }
bias_filler { type: "constant" value: 0 }

} } layer { bottom: "rpn_cls_score" top: "rpn_cls_score_reshape" name: "rpn_cls_score_reshape" type: "Reshape" reshape_param { shape { dim: 0 dim: 2 dim: -1 dim: 0 } } }

========= RoI Proposal ============

layer { name: "rpn_cls_prob" type: "Softmax" bottom: "rpn_cls_score_reshape" top: "rpn_cls_prob" } layer { name: 'rpn_cls_prob_reshape' type: 'Reshape' bottom: 'rpn_cls_prob' top: 'rpn_cls_prob_reshape' reshape_param { shape { dim: 0 dim: 24 dim: -1 dim: 0 } } #原来 18 } layer { name: 'proposal' type: 'Python' bottom: 'rpn_cls_prob_reshape' bottom: 'rpn_bbox_pred' bottom: 'im_info' top: 'rois' python_param { module: 'rpn.proposal_layer' layer: 'ProposalLayer' param_str: "'feat_stride': 16" #16 } }

========= RCNN ============

layer { name: "roi_pool_conv5" type: "ROIPooling" bottom: "conv5" bottom: "rois" top: "roi_pool_conv5" roi_pooling_param { pooled_w: 6 pooled_h: 6 spatial_scale: 0.0625 # 1/16 } } layer { name: "fc6" type: "InnerProduct" bottom: "roi_pool_conv5" top: "fc6" inner_product_param { num_output: 4096 } } layer { name: "relu6" type: "ReLU" bottom: "fc6" top: "fc6" } layer { name: "drop6" type: "Dropout" bottom: "fc6" top: "fc6" dropout_param { dropout_ratio: 0.5 scale_train: false } } layer { name: "fc7" type: "InnerProduct" bottom: "fc6" top: "fc7" inner_product_param { num_output: 4096 } } layer { name: "relu7" type: "ReLU" bottom: "fc7" top: "fc7" } layer { name: "drop7" type: "Dropout" bottom: "fc7" top: "fc7" dropout_param { dropout_ratio: 0.5 scale_train: false } } layer { name: "cls_score_my" type: "InnerProduct" bottom: "fc7" top: "cls_score" inner_product_param { num_output: 4 #按训练集类别改，该值为类别数+1 } } layer { name: "bbox_pred_my" type: "InnerProduct" bottom: "fc7" top: "bbox_pred" inner_product_param { num_output: 16 #按训练集类别改，该值为（类别数+1）*4 } } layer { name: "cls_prob" type: "Softmax" bottom: "cls_score" top: "cls_prob" loss_param { ignore_label: -1 normalize: true } } the number of anchor is 12, (3 ratios, 4 scales)

djdam commented 7 years ago

@jiayandekafei

did you change the anchor_scales for the proposal_layer.py as well?

by the way, I think you shouldn't change the anchor scales in the code. You can configure them in the prototxt like so:

param_str: "{feat_stride: 16, scales: [2, 4, 8, 16]}"

you have to change it for AnchorTargetLayer and ProposalLayer

djdam commented 7 years ago

@jiayandekafei

you're welcome, although you solved it by yourself ;) I prefer not to use qq, don't take it personal, but I don't want to answer questions the whole day if you know what I mean :)

ubenz55555 commented 7 years ago

@djdam @hito0512 I change my anchor to [16 32 64] and keep the ratio as [0.5 1 2 ] ， caused me unable to detect some class in my dataset ， and returns failure. It was fine and gives a result when the anchor is [8 16 32] and [0.5 1 2]， but it fails with some class with [16 32 64] and [0.5 1 2] . Do you know how to solve this problem? def generate_anchors(base_size=32, ratios=[0.5, 1, 2], scales=2**np.arange(4, 7)):

anchor_scales = layer_params.get('scales', (16, 32, 64))

train.prototxt

name: "ZF" layer { name: 'input-data' type: 'Python' top: 'data' top: 'im_info' top: 'gt_boxes' python_param { module: 'roi_data_layer.layer' layer: 'RoIDataLayer' param_str: "'num_classes': 21" } }

========= conv1-conv5 ============

layer { name: "conv1" type: "Convolution" bottom: "data" top: "conv1" param { lr_mult: 1.0 } param { lr_mult: 2.0 } convolution_param { num_output: 96 kernel_size: 7 pad: 3 stride: 2 } } layer { name: "relu1" type: "ReLU" bottom: "conv1" top: "conv1" } layer { name: "norm1" type: "LRN" bottom: "conv1" top: "norm1" lrn_param { local_size: 3 alpha: 0.00005 beta: 0.75 norm_region: WITHIN_CHANNEL engine: CAFFE } } layer { name: "pool1" type: "Pooling" bottom: "norm1" top: "pool1" pooling_param { kernel_size: 3 stride: 2 pad: 1 pool: MAX } } layer { name: "conv2" type: "Convolution" bottom: "pool1" top: "conv2" param { lr_mult: 1.0 } param { lr_mult: 2.0 } convolution_param { num_output: 256 kernel_size: 5 pad: 2 stride: 2 } } layer { name: "relu2" type: "ReLU" bottom: "conv2" top: "conv2" } layer { name: "norm2" type: "LRN" bottom: "conv2" top: "norm2" lrn_param { local_size: 3 alpha: 0.00005 beta: 0.75 norm_region: WITHIN_CHANNEL engine: CAFFE } } layer { name: "pool2" type: "Pooling" bottom: "norm2" top: "pool2" pooling_param { kernel_size: 3 stride: 2 pad: 1 pool: MAX } } layer { name: "conv3" type: "Convolution" bottom: "pool2" top: "conv3" param { lr_mult: 1.0 } param { lr_mult: 2.0 } convolution_param { num_output: 384 kernel_size: 3 pad: 1 stride: 1 } } layer { name: "relu3" type: "ReLU" bottom: "conv3" top: "conv3" } layer { name: "conv4" type: "Convolution" bottom: "conv3" top: "conv4" param { lr_mult: 1.0 } param { lr_mult: 2.0 } convolution_param { num_output: 384 kernel_size: 3 pad: 1 stride: 1 } } layer { name: "relu4" type: "ReLU" bottom: "conv4" top: "conv4" } layer { name: "conv5" type: "Convolution" bottom: "conv4" top: "conv5" param { lr_mult: 1.0 } param { lr_mult: 2.0 } convolution_param { num_output: 256 kernel_size: 3 pad: 1 stride: 1 } } layer { name: "relu5" type: "ReLU" bottom: "conv5" top: "conv5" }

RPN

layer { name: "rpn_conv/3x3" type: "Convolution" bottom: "conv5" top: "rpn/output" param { lr_mult: 1.0 } param { lr_mult: 2.0 } convolution_param { num_output: 256 kernel_size: 3 pad: 1 stride: 1 weight_filler { type: "gaussian" std: 0.01 } bias_filler { type: "constant" value: 0 } } } layer { name: "rpn_relu/3x3" type: "ReLU" bottom: "rpn/output" top: "rpn/output" }

layer { name: "rpn_cls_score" type: "Convolution" bottom: "rpn/output" top: "rpn_cls_score" param { lr_mult: 1.0 } param { lr_mult: 2.0 } convolution_param { num_output: 18 # 2(bg/fg) 9(anchors) kernel_size: 1 pad: 0 stride: 1 weight_filler { type: "gaussian" std: 0.01 } bias_filler { type: "constant" value: 0 } } } layer { name: "rpn_bbox_pred" type: "Convolution" bottom: "rpn/output" top: "rpn_bbox_pred" param { lr_mult: 1.0 } param { lr_mult: 2.0 } convolution_param { num_output: 36 # 4 9(anchors) kernel_size: 1 pad: 0 stride: 1 weight_filler { type: "gaussian" std: 0.01 } bias_filler { type: "constant" value: 0 } } } layer { bottom: "rpn_cls_score" top: "rpn_cls_score_reshape" name: "rpn_cls_score_reshape" type: "Reshape" reshape_param { shape { dim: 0 dim: 2 dim: -1 dim: 0 } } } layer { name: 'rpn-data' type: 'Python' bottom: 'rpn_cls_score' bottom: 'gt_boxes' bottom: 'im_info' bottom: 'data' top: 'rpn_labels' top: 'rpn_bbox_targets' top: 'rpn_bbox_inside_weights' top: 'rpn_bbox_outside_weights' python_param { module: 'rpn.anchor_target_layer' layer: 'AnchorTargetLayer' param_str: "'feat_stride': 16" } } layer { name: "rpn_loss_cls" type: "SoftmaxWithLoss" bottom: "rpn_cls_score_reshape" bottom: "rpn_labels" propagate_down: 1 propagate_down: 0 top: "rpn_cls_loss" loss_weight: 1 loss_param { ignore_label: -1 normalize: true } } layer { name: "rpn_loss_bbox" type: "SmoothL1Loss" bottom: "rpn_bbox_pred" bottom: "rpn_bbox_targets" bottom: 'rpn_bbox_inside_weights' bottom: 'rpn_bbox_outside_weights' top: "rpn_loss_bbox" loss_weight: 1 smooth_l1_loss_param { sigma: 3.0 } }

RoI Proposal

layer { name: "rpn_cls_prob" type: "Softmax" bottom: "rpn_cls_score_reshape" top: "rpn_cls_prob" } layer { name: 'rpn_cls_prob_reshape' type: 'Reshape' bottom: 'rpn_cls_prob' top: 'rpn_cls_prob_reshape' reshape_param { shape { dim: 0 dim: 18 dim: -1 dim: 0 } } } layer { name: 'proposal' type: 'Python' bottom: 'rpn_cls_prob_reshape' bottom: 'rpn_bbox_pred' bottom: 'im_info' top: 'rpn_rois'

top: 'rpn_scores'

python_param { module: 'rpn.proposal_layer' layer: 'ProposalLayer' param_str: "'feat_stride': 16" } }

layer { name: 'roi-data' type: 'Python' bottom: 'rpn_rois' bottom: 'gt_boxes' top: 'rois' top: 'labels' top: 'bbox_targets' top: 'bbox_inside_weights' top: 'bbox_outside_weights' python_param { module: 'rpn.proposal_target_layer' layer: 'ProposalTargetLayer' param_str: "'num_classes': 21" } }

========= RCNN ============

layer { name: "roi_pool_conv5" type: "ROIPooling" bottom: "conv5" bottom: "rois" top: "roi_pool_conv5" roi_pooling_param { pooled_w: 6 pooled_h: 6 spatial_scale: 0.0625 # 1/16 } } layer { name: "fc6" type: "InnerProduct" bottom: "roi_pool_conv5" top: "fc6" param { lr_mult: 1.0 } param { lr_mult: 2.0 } inner_product_param { num_output: 4096 } } layer { name: "relu6" type: "ReLU" bottom: "fc6" top: "fc6" } layer { name: "drop6" type: "Dropout" bottom: "fc6" top: "fc6" dropout_param { dropout_ratio: 0.5 scale_train: false } } layer { name: "fc7" type: "InnerProduct" bottom: "fc6" top: "fc7" param { lr_mult: 1.0 } param { lr_mult: 2.0 } inner_product_param { num_output: 4096 } } layer { name: "relu7" type: "ReLU" bottom: "fc7" top: "fc7" } layer { name: "drop7" type: "Dropout" bottom: "fc7" top: "fc7" dropout_param { dropout_ratio: 0.5 scale_train: false } } layer { name: "cls_score" type: "InnerProduct" bottom: "fc7" top: "cls_score" param { lr_mult: 1.0 } param { lr_mult: 2.0 } inner_product_param { num_output: 21 weight_filler { type: "gaussian" std: 0.01 } bias_filler { type: "constant" value: 0 } } } layer { name: "bbox_pred" type: "InnerProduct" bottom: "fc7" top: "bbox_pred" param { lr_mult: 1.0 } param { lr_mult: 2.0 } inner_product_param { num_output: 84 weight_filler { type: "gaussian" std: 0.001 } bias_filler { type: "constant" value: 0 } } } layer { name: "loss_cls" type: "SoftmaxWithLoss" bottom: "cls_score" bottom: "labels" propagate_down: 1 propagate_down: 0 top: "cls_loss" loss_weight: 1 loss_param { ignore_label: -1 normalize: true } } layer { name: "loss_bbox" type: "SmoothL1Loss" bottom: "bbox_pred" bottom: "bbox_targets" bottom: 'bbox_inside_weights' bottom: 'bbox_outside_weights' top: "bbox_loss" loss_weight: 1 }

wubaorong commented 6 years ago

@djdam Do you meet error in bbox_transform.py py-faster-rcnn/tools/../lib/fast_rcnn/bbox_transform.py:48: RuntimeWarning: overflow encountered in exp pred_w = np.exp(dw) widths[:, np.newaxis] py-faster-rcnn/tools/../lib/fast_rcnn/bbox_transform.py:48: RuntimeWarning: overflow encountered in multiply pred_w = np.exp(dw) widths[:, np.newaxis] py-faster-rcnn/tools/../lib/fast_rcnn/bbox_transform.py:49: RuntimeWarning: overflow encountered in exp pred_h = np.exp(dh) heights[:, np.newaxis] py-faster-rcnn/tools/../lib/fast_rcnn/bbox_transform.py:49: RuntimeWarning: overflow encountered in multiply pred_h = np.exp(dh) heights[:, np.newaxis] /home/wu/faster_rcnn/py-faster-rcnn/tools/../lib/rpn/proposal_layer.py:175: RuntimeWarning: invalid value encountered in greater_equal keep = np.where((ws >= min_size) & (hs >= min_size))[0]

seawei1993 commented 6 years ago

@hito0512 Hello! I changed aspect ratios and followed your method, but when I changed dim: 18 to dim: 24,it didn't work and have some error ,Do you have any idea how to fix it? Thx The error goes like this:

Traceback (most recent call last): File "./tools/train_net.py", line 112, in max_iters=args.max_iters) File "/home/mipl/py-faster-rcnn/tools/../lib/fast_rcnn/train.py", line 160, in train_net model_paths = sw.train_model(max_iters) File "/home/mipl/py-faster-rcnn/tools/../lib/fast_rcnn/train.py", line 101, in train_model self.solver.step(1) File "/home/mipl/py-faster-rcnn/tools/../lib/rpn/proposal_layer.py", line 122, in forward proposals = bbox_transform_inv(anchors, bbox_deltas) File "/home/mipl/py-faster-rcnn/tools/../lib/fast_rcnn/bbox_transform.py", line 46, in bbox_transform_inv pred_ctr_x = dx * widths[:, np.newaxis] + ctr_x[:, np.newaxis] ValueError: operands could not be broadcast together with shapes (27648,1) (20736,1)

DongXingshuai commented 6 years ago

@seawei1993 Hi, I also met this problem, have you found the reason or solution? Thanks!

hito0512 commented 6 years ago

@seawei1993，@DongXingshuai
sorry, The modified file (.prototext)is not read ,Modify the path of prototext

hito0512 commented 6 years ago

@ubenz55555 sorry, The modified file (.prototext)is not read ,Modify the path of prototext

rbgirshick / py-faster-rcnn

change anchor size #600

num_output: 18 # 2(bg/fg) * 9(anchors)

num_output: 36 # 4 * 9(anchors)

========= conv1-conv5 ============

================ #将此池化层删除 ===============

===========================================

========= RPN ============

num_output: 40 # 2(bg/fg) * 20(anchors)

num_output: 80 # 4 * 20(anchors)

========= RoI Proposal ============

========= RCNN ============

========= conv1-conv5 ============

=========================================

=========================================

========= RPN ============

====================================

num_output: 40 # 2(bg/fg) * 20(anchors)

num_output: 80 # 4 * 20(anchors)

========= RoI Proposal ============

========= RCNN ============

========= conv1-conv5 ============

top: 'rpn_scores'

========= RCNN ============