Forums - Failure of SNPE dlc conversion of faster rcnn tensorflow model

1 post / 0 new
or Register
Failure of SNPE dlc conversion of faster rcnn tensorflow model
virtuoso.wu
Join Date: 7 Feb 21
Posts: 1
Posted: Tue, 2021-03-23 10:11
Top

Hi,

 snpe-1.48.0.2554

Python 3.6

tensorflow 1.5

faster rcnn tensorflow models:

faster_rcnn_resnet101_kitti: http://download.tensorflow.org/models/object_detection/faster_rcnn_resne...

faster_rcnn_inception_resnet_v2_atrous_oidv2: http://download.tensorflow.org/models/object_detection/faster_rcnn_incep...

I tried to convert both models but they all failed with the error below:

    ==============================================================
    Building Network
    ==============================================================
2021-03-23 10:08:51,275 - 172 - ERROR - Conversion FAILED: Cannot find bboxes or scores
Traceback (most recent call last):
  File "/home/hhwu/project/snpe-1.48.0.2554/bin/x86_64-linux-clang/snpe-tensorflow-to-dlc", line 49, in main
    ir_graph = converter.convert()
  File "/home/hhwu/project/snpe-1.48.0.2554/lib/python/qti/aisw/converters/tensorflow/tf_to_ir.py", line 326, in convert
    self._convert_layers()
  File "/home/hhwu/project/snpe-1.48.0.2554/lib/python/qti/aisw/converters/tensorflow/tf_to_ir.py", line 359, in _convert_layers
    descriptors = self._resolve_descriptors_from_nodes(graph_ops)
  File "/home/hhwu/project/snpe-1.48.0.2554/lib/python/qti/aisw/converters/tensorflow/tf_to_ir.py", line 497, in _resolve_descriptors_from_nodes
    resolved_descriptors = resolver.resolve_layer(graph_matcher, self._graph_helper)
  File "/home/hhwu/project/snpe-1.48.0.2554/lib/python/qti/aisw/converters/tensorflow/layers/non_max_suppression.py", line 218, in resolve_layer
    self._resolve_for_gather_layer(graph_matcher, graph_helper, nms_descriptor)
  File "/home/hhwu/project/snpe-1.48.0.2554/lib/python/qti/aisw/converters/tensorflow/layers/non_max_suppression.py", line 172, in _resolve_for_gather_layer
    raise ConverterError('Cannot find bboxes or scores')
qti.aisw.converters.tensorflow.util.ConverterError: Cannot find bboxes or scores
 
Have you guys tried to convert faster rcnn before?
Would you like to share how to do it? Thanks.
  • Up0
  • Down0

Opinions expressed in the content posted here are the personal opinions of the original authors, and do not necessarily reflect those of Qualcomm Incorporated or its subsidiaries (“Qualcomm”). The content is provided for informational purposes only and is not meant to be an endorsement or representation by Qualcomm or any other party. This site may also provide links or references to non-Qualcomm sites and resources. Qualcomm makes no representations, warranties, or other commitments whatsoever about any non-Qualcomm sites or third-party resources that may be referenced, accessible from, or linked to this site.