can't train my datasets

[mp168]

  File "D:/mp_code/smog_faster_rcnn/train.py", line 154, in train
    blobs = self.data_layer.forward()
  File "D:\mp_code\smog_faster_rcnn\lib\layer_utils\roi_data_layer.py", line 75, in forward
    blobs = self._get_next_minibatch()
  File "D:\mp_code\smog_faster_rcnn\lib\layer_utils\roi_data_layer.py", line 71, in _get_next_minibatch
    return get_minibatch(minibatch_db, self._num_classes)
  File "D:\mp_code\smog_faster_rcnn\lib\utils\minibatch.py", line 36, in get_minibatch
    im_blob, im_scales = _get_image_blob(roidb, random_scale_inds)
  File "D:\mp_code\smog_faster_rcnn\lib\utils\minibatch.py", line 70, in _get_image_blob
    im = cv2.imread(roidb[i]['image'])
KeyError: 'image'

I had seen the same questions, but i still can't solve it, I try to my data as yours, can you help me?

[morpheusthewhite]

As I answered in the previous question this depends on you dataset, probably it is not good for this net.

The only way to solve it is to add the missing key in each xml file

[mp168]

<?xml version="1.0"?>

-

VOC2007 000001.jpg

-

The VOC2007 Database PASCAL VOC2007 flickr

-

Fried Camels Jinky the Fruit Bat

-

704 576 3

0

-

this is my xml. and i only recognite one class,I can't find my datasets where is wrong??!!

[morpheusthewhite]

Did you open it with a text editor (notepad if Windows or kate on linux) to paste the content here?

If not, do it and paste the content here

[mp168]

VOC2007 000007.jpg The VOC2007 Database PASCAL VOC2007 flickr Fried Camels Jinky the Fruit Bat 704 576 3 0

[morpheusthewhite]

Again, the keys are missing. I need it in the form

<key1>
    <key2>value</key2>
etc.
</key1>

[mp168]

<annotation>
    <folder>VOC2007</folder>
    <filename>000007.jpg</filename>
    <source>
        <database>The VOC2007 Database</database>
        <annotation>PASCAL VOC2007</annotation>
        <image>flickr</image>
    </source>
    <owner>
        <flickrid>Fried Camels</flickrid>
        <name>Jinky the Fruit Bat</name>
    </owner>
    <size>
        <width>704</width>
        <height>576</height>
        <depth>3</depth>
    </size>
    <segmented>0</segmented>
    <object>
        <name>smog</name>
        <pose>Unspecified</pose>
        <truncated>0</truncated>
        <difficult>0</difficult>
        <bndbox>
            <xmin>176</xmin>
            <ymin>260</ymin>
            <xmax>209</xmax>
            <ymax>293</ymax>
        </bndbox>
    </object>
    <object>
        <name>smog</name>
        <pose>Unspecified</pose>
        <truncated>0</truncated>
        <difficult>0</difficult>
        <bndbox>
            <xmin>221</xmin>
            <ymin>260</ymin>
            <xmax>250</xmax>
            <ymax>296</ymax>
        </bndbox>
    </object>
</annotation>

[morpheusthewhite]

The xml file is correct, so it should work

Make sure you placed the files in the right location (see https://github.com/dBeker/Faster-RCNN-TensorFlow-Python3/issues/82)

[lw17809213284]

when i run the train.py,TypeError: argument of type 'NoneType' is not iterable

[BIG-DD]

  File "D:/mp_code/smog_faster_rcnn/train.py", line 154, in train
    blobs = self.data_layer.forward()
  File "D:\mp_code\smog_faster_rcnn\lib\layer_utils\roi_data_layer.py", line 75, in forward
    blobs = self._get_next_minibatch()
  File "D:\mp_code\smog_faster_rcnn\lib\layer_utils\roi_data_layer.py", line 71, in _get_next_minibatch
    return get_minibatch(minibatch_db, self._num_classes)
  File "D:\mp_code\smog_faster_rcnn\lib\utils\minibatch.py", line 36, in get_minibatch
    im_blob, im_scales = _get_image_blob(roidb, random_scale_inds)
  File "D:\mp_code\smog_faster_rcnn\lib\utils\minibatch.py", line 70, in _get_image_blob
    im = cv2.imread(roidb[i]['image'])
KeyError: 'image'

<annotation>
    <folder>VOC2007</folder>
    <filename>000007.jpg</filename>
    <source>
        <database>The VOC2007 Database</database>
      <annotation>PASCAL VOC2007</annotation>
      <image>flickr</image>
    </source>
  <owner>
      <flickrid>Fried Camels</flickrid>
      <name>Jinky the Fruit Bat</name>
  </owner>
    <size>
        <width>704</width>
        <height>576</height>
        <depth>3</depth>
    </size>
    <segmented>0</segmented>
    <object>
        <name>smog</name>
        <pose>Unspecified</pose>
        <truncated>0</truncated>
        <difficult>0</difficult>
        <bndbox>
            <xmin>176</xmin>
            <ymin>260</ymin>
            <xmax>209</xmax>
            <ymax>293</ymax>
        </bndbox>
    </object>
    <object>
        <name>smog</name>
        <pose>Unspecified</pose>
        <truncated>0</truncated>
        <difficult>0</difficult>
        <bndbox>
            <xmin>221</xmin>
            <ymin>260</ymin>
            <xmax>250</xmax>
            <ymax>296</ymax>
        </bndbox>
    </object>
</annotation>

  File "D:/mp_code/smog_faster_rcnn/train.py", line 154, in train
    blobs = self.data_layer.forward()
  File "D:\mp_code\smog_faster_rcnn\lib\layer_utils\roi_data_layer.py", line 75, in forward
    blobs = self._get_next_minibatch()
  File "D:\mp_code\smog_faster_rcnn\lib\layer_utils\roi_data_layer.py", line 71, in _get_next_minibatch
    return get_minibatch(minibatch_db, self._num_classes)
  File "D:\mp_code\smog_faster_rcnn\lib\utils\minibatch.py", line 36, in get_minibatch
    im_blob, im_scales = _get_image_blob(roidb, random_scale_inds)
  File "D:\mp_code\smog_faster_rcnn\lib\utils\minibatch.py", line 70, in _get_image_blob
    im = cv2.imread(roidb[i]['image'])
KeyError: 'image'

I had seen the same questions, but i still can't solve it, I try to my data as yours, can you help me?

Has your problem been solved? I encountered this problem too.