使用external_converter_v2目录下的converter.py转换mobilenet-ssd的caffemodle模型失败

[wzzju]

config.yaml的文件内容如下:

OPTIONS:
    Framework: CAFFE
    SavePath: ./output
    ResultName: mobilenet-ssd
    Config:
        LaunchBoard: ON
        Server:
            ip: 0.0.0.0
            port: 8888
        OptimizedGraph: 
            enable: OFF
            path: /home/public/CLionProjects/Anakin/tools/external_converter_v2/anakin_optimized/mobilenet.anakin.bin.saved
    LOGGER:
        LogToPath: ./log/
        WithColor: ON 

TARGET:
    CAFFE:
        # path to proto files
        ProtoPaths:
            - /home/public/CLionProjects/Anakin/tools/external_converter_v2/caffe.proto
        PrototxtPath: /home/public/CLionProjects/Anakin/tools/external_converter_v2/mobilenet-ssd/MobileNetSSD_deploy.prototxt
        ModelPath: /home/public/CLionProjects/Anakin/tools/external_converter_v2/mobilenet-ssd/MobileNetSSD_deploy.caffemodel

其中caffe.proto文件使用来源为:https://github.com/weiliu89/caffe/blob/ssd/src/caffe/proto/caffe.proto MobileNet-SSD模型来源为:https://github.com/chuanqi305/MobileNet-SSD

运行python converter.py命令,执行过程中出现如下错误:

INF | 2018-09-12 19:56:12.527134 | parser_caffe.py:539 _Parsing_new() ]    `--[ Match ]Parsing [Permute:    conv11_mbox_conf_perm] 
INF | 2018-09-12 19:56:12.527407 | parser_caffe.py:524 _Parsing_new() ]   Dectect [Flatten: conv11_mbox_conf_flat] 
INF | 2018-09-12 19:56:12.527663 | parser_caffe.py:539 _Parsing_new() ]    `--[ Match ]Parsing [Flatten:    conv11_mbox_conf_flat] 
INF | 2018-09-12 19:56:12.527922 | parser_caffe.py:524 _Parsing_new() ]   Dectect [PriorBox:    conv11_mbox_priorbox] 
INF | 2018-09-12 19:56:12.528170 | parser_caffe.py:539 _Parsing_new() ]    `--[ Match ]Parsing [PriorBox:   conv11_mbox_priorbox] 
Traceback (most recent call last):
  File "converter.py", line 20, in <module>
    graph = Graph(config)
  File "/home/public/CLionProjects/Anakin/tools/external_converter_v2/parser/graph.py", line 39, in __init__
    self.graph_io = self.parser()
  File "/home/public/CLionProjects/Anakin/tools/external_converter_v2/parser/caffe/parser_caffe.py", line 34, in __call__
    return self._Parsing_new()
  File "/home/public/CLionProjects/Anakin/tools/external_converter_v2/parser/caffe/parser_caffe.py", line 570, in _Parsing_new
    CAFFE_LAYER_PARSER[source_layer_type](nodeIO, rlayer, tensors, opIO) # call parser automatically
  File "/home/public/CLionProjects/Anakin/tools/external_converter_v2/parser/operations/op.py", line 42, in __call__
    return self.parser(args)
  File "/home/public/CLionProjects/Anakin/tools/external_converter_v2/parser/caffe/caffe_layer_param_transmit.py", line 39, in warpper_args
    Parser(args)
  File "/home/public/CLionProjects/Anakin/tools/external_converter_v2/parser/caffe/caffe_layer_param_transmit.py", line 1081, in Parser_priorbox
    OpsRegister()["PriorBox"].fixed_size = list(prior_box_param.fixed_size)
AttributeError: 'PriorBoxParameter' object has no attribute 'fixed_size'

请求帮助,非常感谢!

[xyoungli]

caffe.proto 中，在priorbox参数添加3个参数： repeated float fixed_size = 20; repeated float fixed_ratio = 21; repeated float density = 22;

[wzzju]

@xyoungli 谢谢。

[ljayx]

@xyoungli 你好，我加了这三个参数到caffe.proto里：

message PriorBoxParameter {
  // Encode/decode type.
  enum CodeType {
    CORNER = 1;
    CENTER_SIZE = 2;
    CORNER_SIZE = 3;
  }
  // Minimum box size (in pixels). Required!
  repeated float min_size = 1;
  // Maximum box size (in pixels). Required!
  repeated float max_size = 2;
  // Various of aspect ratios. Duplicate ratios will be ignored.
  // If none is provided, we use default ratio 1.
  repeated float aspect_ratio = 3;
  // If true, will flip each aspect ratio.
  // For example, if there is aspect ratio "r",
  // we will generate aspect ratio "1.0/r" as well.
  optional bool flip = 4 [default = true];
  // If true, will clip the prior so that it is within [0, 1]
  optional bool clip = 5 [default = false];
  // Variance for adjusting the prior bboxes.
  repeated float variance = 6;
  // By default, we calculate img_height, img_width, step_x, step_y based on
  // bottom[0] (feat) and bottom[1] (img). Unless these values are explicitely
  // provided.
  // Explicitly provide the img_size.
  optional uint32 img_size = 7;
  // Either img_size or img_h/img_w should be specified; not both.
  optional uint32 img_h = 8;
  optional uint32 img_w = 9;

  // Explicitly provide the step size.
  optional float step = 10;
  // Either step or step_h/step_w should be specified; not both.
  optional float step_h = 11;
  optional float step_w = 12;

  // Offset to the top left corner of each cell.
  optional float offset = 13 [default = 0.5];

  // lj
  repeated float fixed_size = 20;
  repeated float fixed_ratio = 21;
  repeated float density = 22;
}

但是我又遇到了另一个问题：

Traceback (most recent call last):
  File "/Users/xxx/Anakin/tools/external_converter_v2/converter.py", line 20, in <module>
    graph = Graph(config)
  File "/Users/xxx/Anakin/tools/external_converter_v2/parser/graph.py", line 23, in __init__
    from caffe import CaffeParser
  File "/Users/xxx/Anakin/tools/external_converter_v2/parser/caffe/__init__.py", line 5, in <module>
    from parser_caffe import *
  File "/Users/xxx/Anakin/tools/external_converter_v2/parser/caffe/parser_caffe.py", line 9, in <module>
    from caffe_helper import *
  File "/Users/xxx/Anakin/tools/external_converter_v2/parser/caffe/caffe_helper.py", line 6, in <module>
    from ..pbs import *
  File "/Users/xxx/Anakin/tools/external_converter_v2/parser/pbs/__init__.py", line 13, in <module>
    m = __import__(module[:-3], locals(), globals())
  File "/Users/xxx/Anakin/tools/external_converter_v2/parser/pbs/caffe_pb2.py", line 23, in <module>
    fle_images\x18@ \x01(\x08:\x05\x66\x61lse\x12\x15\n\nconcat_dim\x18\x41 \x01(\r:\x01\x31\x12\x36\n\x11hdf5_output_param\x18\xe9\x07 \x01(\x0b\x32\x1a.caffe.HDF5OutputParameter\".\n\nPoolMethod\x12\x07\n\x03MAX\x10\x00\x12\x07\n\x03\x41VE\x10\x01\x12\x0e\n\nSTOCHASTIC\x10\x02\"W\n\x0ePReLUParameter\x12&\n\x06\x66iller\x18\x01 \x01(\x0b\x32\x16.caffe.FillerParameter\x12\x1d\n\x0e\x63hannel_shared\x18\x02 \x01(\x08:\x05\x66\x61lse*\x1c\n\x05Phase\x12\t\n\x05TRAIN\x10\x00\x12\x08\n\x04TEST\x10\x01')
  File "/Users/xxx/anaconda2/envs/vf/lib/python2.7/site-packages/google/protobuf/descriptor.py", line 878, in __new__
    return _message.default_pool.AddSerializedFile(serialized_pb)
TypeError: Couldn't build proto file into descriptor pool!
Invalid proto descriptor for file "caffe.proto":
  caffe.proto: A file with this name is already in the pool.

请问应该怎么解决呢？ Thanks in advance!

[ljayx]

Anakin branch: developing (already up to date) protobuf version: 3.6.0

[Shixiaowei02]

你的分支是 developing。这个问题是 protoc 尝试导入多个相同文件名的 proto 引起的冲突。你的描述看起来是 caffe.proto 冲突了。最新版本的 parser 会在运行之前清理 proto 缓存，请切换到 master 分支试试。@Knownx

[ljayx]

@Shixiaowei02 切换到master分支并更新代码后问题解决，谢谢！