Forums - Cannot use GPU to build Batch input mobilenet ssd network

4 posts / 0 new
or Register
Last post
Cannot use GPU to build Batch input mobilenet ssd network
Jacob_Ke
Join Date: 24 Aug 17
Posts: 10
Posted: Wed, 2018-08-01 01:19
Top

Hi,

I'm try batch for mobilenet ssd on android

but output error when i build network

Model source:

wget http://download.tensorflow.org/models/object_detection/ssd_mobilenet_v1_coco_2018_01_28.tar.gz
 

Convert  model comand:

snpe-tensorflow-to-dlc --graph ssd_mobilenet_v1_coco_2018_01_28/frozen_inference_graph.pb -i Preprocessor/sub 2,300,300,3 --out_node detection_classes --out_node detection_boxes --out_node detection_scores --dlc mobilenet_ssd.dlc --allow_unconsumed_nodes
 
Builder info on android:
final SNPE.NeuralNetworkBuilder builder = new SNPE.NeuralNetworkBuilder(mActivity.getApplication())
                    .setDebugEnabled(false)
                    .setOutputLayers("Postprocessor/BatchMultiClassNonMaxSuppression")
                    .setRuntimeOrder(NeuralNetwork.Runtime.GPU)
                    .setCpuFallbackEnabled(true)
network = builder.build();
                  
snpe sdk version: 1.17.0
 
error message:
E/LoadNetworkTask: Unable to create network! Cause: error_code=811; error_message=GPU tensor dimensions are invalid. Tensor dimension mismatch; error_component=GPU Runtime; line_no=103; thread_id=-399001312
    java.lang.IllegalStateException: Unable to create network! Cause: error_code=811; error_message=GPU tensor dimensions are invalid. Tensor dimension mismatch; error_component=GPU Runtime; line_no=103; thread_id=-399001312
        at com.qualcomm.qti.snpe.internal.NativeNetwork.nativeInitFromFile(Native Method)
 
 
try case:
1. Runtime.CPU is OK, but GPU is failed
2. Two different phones(zenfone 5Z, zenfone 3) and the processer S845 and S625 have the same situation
3. The comand "snpe-tensorflow-to-dlc --graph ssd_mobilenet_v1_coco_2018_01_28/frozen_inference_graph.pb -i Preprocessor/sub 1,300,300,3 --out_node detection_classes --out_node detection_boxes --out_node detection_scores --dlc mobilenet_ssd.dlc --allow_unconsumed_nodes" is work fine!
 
 
How can i fix it?
 
Thanks
  • Up0
  • Down0
Re: Cannot use GPU to build Batch input mobilenet ssd network
Enrico Ros
Join Date: 14 Jul 17
Posts: 10
Posted: Wed, 2018-08-01 11:56
Top

Would i Preprocessor/sub 1,300,300,3 (changed 2 to 1) make it work? with single and multiple images?

  • Up0
  • Down0
Re: Cannot use GPU to build Batch input mobilenet ssd network
Jacob_Ke
Join Date: 24 Aug 17
Posts: 10
Posted: Wed, 2018-08-01 23:46
Top

Hi Enrico Ros,

I try Converting model by 1,300,300,3 is OK and builder build network also work fine!

But I cannot input multiple images, mNeuralNetwork.getInputTensorsShapes().get("Preprocessor/sub:0") is {1,300,300,3}

 

  • Up0
  • Down0
Re: Cannot use GPU to build Batch input mobilenet ssd network
zf.africa
Join Date: 15 Jun 17
Posts: 51
Posted: Wed, 2018-08-08 00:13
Top

Hi Enrico,

Does user buffer mode supports batch input now?

I saw that batch input is supported with ITensor input, but user buffer mode is not mentioned in document.

Any comments are appreciated, thanks.

  • Up0
  • Down0
or Register

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.