- input: "data"
- input_dim: 1
- input_dim: 3
- input_dim: 16
- input_dim: 66
- layer {
- name: "conv1"
- type: "Convolution"
- bottom: "data"
- top: "conv1"
- convolution_param {
- num_output: 10
- bias_term: true
- pad: 0
- kernel_size: 3
- stride: 1
- }
- }
- layer {
- name: "relu1"
- type: "ReLU"
- bottom: "conv1"
- top: "conv1"
- }
- layer {
- name: "max_pooling2d_3"
- type: "Pooling"
- bottom: "conv1"
- top: "max_pooling2d_3"
- pooling_param {
- pool: MAX
- kernel_size: 2
- stride: 2
- pad: 0
- }
- }
- layer {
- name: "conv2"
- type: "Convolution"
- bottom: "max_pooling2d_3"
- top: "conv2"
- convolution_param {
- num_output: 16
- bias_term: true
- pad: 0
- kernel_size: 3
- stride: 1
- }
- }
- layer {
- name: "relu2"
- type: "ReLU"
- bottom: "conv2"
- top: "conv2"
- }
- layer {
- name: "conv3"
- type: "Convolution"
- bottom: "conv2"
- top: "conv3"
- convolution_param {
- num_output: 32
- bias_term: true
- pad: 0
- kernel_size: 3
- stride: 1
- }
- }
- layer {
- name: "relu3"
- type: "ReLU"
- bottom: "conv3"
- top: "conv3"
- }
- layer {
- name: "flatten_2"
- type: "Flatten"
- bottom: "conv3"
- top: "flatten_2"
- }
- layer {
- name: "dense"
- type: "InnerProduct"
- bottom: "flatten_2"
- top: "dense"
- inner_product_param {
- num_output: 2
- }
- }
- layer {
- name: "relu4"
- type: "ReLU"
- bottom: "dense"
- top: "dense"
- }
复制代码
我改成- # input: "data"
- # input_dim: 1
- # input_dim: 3
- # input_dim: 16
- # input_dim: 66
- layer {
- name: "data"
- type: "Input"
- top: "data"
- input_param { shape: { dim: 1 dim: 3 dim: 16 dim: 66 } }
- }
- layer {
- name: "conv1"
- type: "Convolution"
- bottom: "data"
- top: "conv1"
- convolution_param {
- num_output: 10
- bias_term: true
- pad: 0
- kernel_size: 3
- stride: 1
- }
- }
- layer {
- name: "relu1"
- type: "ReLU"
- bottom: "conv1"
- top: "conv1"
- }
- layer {
- name: "max_pooling2d_3"
- type: "Pooling"
- bottom: "conv1"
- top: "max_pooling2d_3"
- pooling_param {
- pool: MAX
- kernel_size: 2
- stride: 2
- pad: 0
- }
- }
- layer {
- name: "conv2"
- type: "Convolution"
- bottom: "max_pooling2d_3"
- top: "conv2"
- convolution_param {
- num_output: 16
- bias_term: true
- pad: 0
- kernel_size: 3
- stride: 1
- }
- }
- layer {
- name: "relu2"
- type: "ReLU"
- bottom: "conv2"
- top: "conv2"
- }
- layer {
- name: "conv3"
- type: "Convolution"
- bottom: "conv2"
- top: "conv3"
- convolution_param {
- num_output: 32
- bias_term: true
- pad: 0
- kernel_size: 3
- stride: 1
- }
- }
- layer {
- name: "relu3"
- type: "ReLU"
- bottom: "conv3"
- top: "conv3"
- }
- layer {
- name: "flatten_2"
- type: "Flatten"
- bottom: "conv3"
- top: "flatten_2"
- }
- layer {
- name: "dense"
- type: "InnerProduct"
- bottom: "flatten_2"
- top: "dense"
- inner_product_param {
- num_output: 2
- }
- }
- layer {
- name: "relu4"
- type: "ReLU"
- bottom: "dense"
- top: "dense"
- }
复制代码
模型在加载阶段rknn.load_caffe就出错了,报错信息如下所示- --> config model
- done
- --> Loading model
- E Catch exception when loading caffe model: ./HorizonalFinemapping.prototxt!
- E Traceback (most recent call last):
- E File "rknn\api\rknn_base.py", line 441, in rknn.api.rknn_base.RKNNBase.load_caffe
- E File "rknn\base\RKNNlib\converter\caffeloader.py", line 1042, in rknn.base.RKNNlib.converter.caffeloader.CaffeLoader.load_blobs
- E File "rknn\base\RKNNlib\converter\caffeloader.py", line 935, in rknn.base.RKNNlib.converter.caffeloader.CaffeLoader.parse_blobs
- E File "rknn\base\RKNNlib\converter\caffeloader.py", line 358, in rknn.base.RKNNlib.converter.caffeloader.proc_blobs_convolution
- E File "G:\ProgramData\Anaconda3\envs\rknn\lib\site-packages\numpy\core\fromnumeric.py", line 292, in reshape
- E return _wrapfunc(a, 'reshape', newshape, order=order)
- E File "G:\ProgramData\Anaconda3\envs\rknn\lib\site-packages\numpy\core\fromnumeric.py", line 56, in _wrapfunc
- E return getattr(obj, method)(*args, **kwds)
- E ValueError: cannot reshape array of size 0 into shape (10,3,3,3)
- Load model failed! Ret = -1
复制代码
模型文件地址是https://github.com/zeusees/Hyper ... nception.caffemodel- D Save log info to: ./caffe.log
- --> config model
- done
- --> Loading model
- I Set caffe proto to caffe
- I Load caffe model ./HorizonalFinemapping.prototxt
- I Parsing net parameters ...
- D import clients finished
- I Parsing layer parameters ...
- D Convert layer image
- D Convert layer conv1
- D Convert layer relu1
- D Convert layer max_pooling2d_3
- D Convert layer conv2
- D Convert layer relu2
- D Convert layer conv3
- D Convert layer relu3
- D Convert layer flatten_2
- D Convert layer dense
- D Convert layer relu4
- I Parsing connections ...
- D Connect: image_0,0 to conv1_1,0
- D Connect: conv1_1,0 to relu1_2,0
- D Connect: relu1_2,0 to max_pooling2d_3_3,0
- D Connect: max_pooling2d_3_3,0 to conv2_4,0
- D Connect: conv2_4,0 to relu2_5,0
- D Connect: relu2_5,0 to conv3_6,0
- D Connect: conv3_6,0 to relu3_7,0
- D Connect: relu3_7,0 to flatten_2_8,0
- D Connect: flatten_2_8,0 to dense_9,0
- D Connect: dense_9,0 to relu4_10,0
- D Connect: relu4_10,0 to output_11,0,
- I Load net complete.
- 2020-06-29 16:50:26.078880: I tensorflow/core/platform/cpu_feature_guard.cc:141] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2
- D Process image_0 ...
- D RKNN output shape(input): (0 16 66 3)
- D Process conv1_1 ...
- D RKNN output shape(convolution): (0 14 64 10)
- D Process relu1_2 ...
- D RKNN output shape(relu): (0 14 64 10)
- D Process max_pooling2d_3_3 ...
- D RKNN output shape(pooling): (0 7 32 10)
- D Process conv2_4 ...
- D RKNN output shape(convolution): (0 5 30 16)
- D Process relu2_5 ...
- D RKNN output shape(relu): (0 5 30 16)
- D Process conv3_6 ...
- D RKNN output shape(convolution): (0 3 28 32)
- D Process relu3_7 ...
- D RKNN output shape(relu): (0 3 28 32)
- D Process flatten_2_8 ...
- D RKNN output shape(flatten): (0 2688)
- D Process dense_9 ...
- D RKNN output shape(fullconnect): (0 2)
- D Process relu4_10 ...
- D RKNN output shape(relu): (0 2)
- D Process output_11 ...
- D RKNN output shape(output): (0 2)
- I Build complete.
- I Load blobs from caffe model ./HorizonalFinemapping.caffemodel
- I Check if ./HorizonalFinemapping.caffemodel contains 'type' value
- I Parsing net blobs ...
- D Load blobs of conv1
- E Catch exception when loading caffe model: ./HorizonalFinemapping.prototxt!
- E Traceback (most recent call last):
- E File "rknn\api\rknn_base.py", line 441, in rknn.api.rknn_base.RKNNBase.load_caffe
- E File "rknn\base\RKNNlib\converter\caffeloader.py", line 1042, in rknn.base.RKNNlib.converter.caffeloader.CaffeLoader.load_blobs
- E File "rknn\base\RKNNlib\converter\caffeloader.py", line 935, in rknn.base.RKNNlib.converter.caffeloader.CaffeLoader.parse_blobs
- E File "rknn\base\RKNNlib\converter\caffeloader.py", line 358, in rknn.base.RKNNlib.converter.caffeloader.proc_blobs_convolution
- E File "G:\ProgramData\Anaconda3\envs\rknn\lib\site-packages\numpy\core\fromnumeric.py", line 292, in reshape
- E return _wrapfunc(a, 'reshape', newshape, order=order)
- E File "G:\ProgramData\Anaconda3\envs\rknn\lib\site-packages\numpy\core\fromnumeric.py", line 56, in _wrapfunc
- E return getattr(obj, method)(*args, **kwds)
- E ValueError: cannot reshape array of size 0 into shape (10,3,3,3)
- Load model failed! Ret = -1
复制代码
欢迎光临 Toybrick (https://t.rock-chips.com/) | Powered by Discuz! X3.3 |