MobileNet SSD V2 converted tflite output

[9friday]

System information

• Have I written custom code (as opposed to using a stock example script provided in TensorFlow): N/A
• OS Platform and Distribution (e.g., Linux Ubuntu 16.04): Linux Ubuntu 18.04
• Mobile device (e.g. iPhone 8, Pixel 2, Samsung Galaxy) if the issue happens on mobile device: QUALCOMM MSM8996 for arm64(Development Board), motorola one power
• TensorFlow installed from (source or binary): binary
• TensorFlow version (use command below): 1.15
• Python version: 3.6.9
• Bazel version (if compiling from source): 2.0.0
• GCC/Compiler version (if compiling from source): 8.3.0
• CUDA/cuDNN version: N/A
• GPU model and memory: N/A
• Exact command to reproduce:

Describe the problem

I am working on object detection application on android using the TensorflowLite C++API. When I convert ssd_mobilenet_v2_coco_2018_03_29 model to tflite, output of converted tflite model differs from output of originally provided .pb file on the same test images.

Converting the .pb file to .tflite

Terminal Command(I used this link as example.):

python3 object_detection/export_tflite_ssd_graph.py --pipeline_config_path=/content/ssd_mobilenet_v2_coco_2018_03_29/pipeline.config --trained_checkpoint_prefix=/content/ssd_mobilenet_v2_coco_2018_03_29/model.ckpt --output_directory=/tmp/tflite_graph --add_postprocessing_op=true

The above command threw an error:

google.protobuf.text_format.ParseError: 109:7 : Message type "object_detection.protos.SsdFeatureExtractor" has no field named "batch_norm_trainable"

The error was resolved after following the steps suggested here.

But, the detection output of the converted tflite model differs from that of the originaly provided .pb file.

Output Images:

Using .pb (and .pbtxt) files:

Using converted .tflite file:

Any help and guidance would be appreciated. Cheers:)

[gadagashwini]

Hi @9friday, Follow the instructions mentioned on the tutorial https://github.com/tensorflow/models/blob/master/research/object_detection/colab_tutorials/eager_few_shot_od_training_tflite.ipynb. Thank you!

[google-ml-butler[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you.

[google-ml-butler[bot]]

Closing as stale. Please reopen if you'd like to work on this further.

[google-ml-butler[bot]]

Are you satisfied with the resolution of your issue? Yes No