Latest docker image: error: unrecognized arguments

[MaHoef]

Hello Pinto,
I am following the tutorial form here:
https://github.com/geaxgx/openvino_hand_tracker
Can it be that in the latest docker version the API has changed?

This is what I do:
marco@CHROTZ:/tmp/depthai_hand_tracker$ docker run --gpus all -it --rm
-v pwd:/workspace/resources
-e LOCAL_UID=$(id -u $USER)
-e LOCAL_GID=$(id -g $USER)
pinto0309/tflite2tensorflow:latest bash

NVIDIA Release 20.09 (build 15985252)

NVIDIA TensorRT 7.1.3 (c) 2016-2020, NVIDIA CORPORATION. All rights reserved.
Container image (c) 2020, NVIDIA CORPORATION. All rights reserved.

https://developer.nvidia.com/tensorrt

To install Python sample dependencies, run /opt/tensorrt/python/python_setup.sh

To install open source parsers, plugins, and samples, run /opt/tensorrt/install_opensource.sh. See https://github.com/NVIDIA/TensorRT for more information.

error: XDG_RUNTIME_DIR not set in the environment.
[setupvars.sh] OpenVINO environment initialized

wget https://github.com/google/mediapipe/blob/master/mediapipe/modules/hand_landmark/hand_landmark.tflite
wget https://github.com/google/mediapipe/blob/master/mediapipe/modules/palm_detection/palm_detection.tflite

tflite2tensorflow --model_path ./hand_landmark.tflite --model_output_path palm_detection --flatc_path /home/user/flatc --schema_path /home/user/schema.fbs --output_pb True
tflite2tensorflow: error: unrecognized arguments: True

tflite2tensorflow --model_path ./hand_landmark.tflite --model_output_path palm_detection --flatc_path /home/user/flatc --schema_path /home/user/schema.fbs --output_pb
output json command = /home/user/flatc -t --strict-json --defaults-json -o . /home/user/schema.fbs -- ./hand_landmark.tflite
/home/user/flatc: error: binary "./hand_landmark.tflite" does not have expected file_identifier "TFL3", use --raw-binary to read this file anyway.

Thanks
Marco

[PINTO0309]

You have issued a Pull Request to the repository you are referring to.

geaxgx/openvino_hand_tracker#1

[MaHoef]

Hi Pinto,
thanks for the quick replay.
I have seen the command line change.
However, I still can't convert the google models to TF as the tool claims: does not have expected file_identifier "TFL3"
I don't need the openvino outputs. My intention is to use a TF float model, convert it to the h5 format and use the Xilinx compiler with it.

[PINTO0309]

I'm going to bed now since it's late at night today, but I'll check tomorrow. I'm not sure, but updating to tf-nightly might work.

[PINTO0309]

@MaHoef
To generate .h5, which is very old and officially deprecated, a special procedure is required; to generate saved_model, it can be easily converted without any particular error. I don't know the specifics of Xilinx compiler, but I would not recommend using .h5.

$ docker run -it --rm \
  -v `pwd`:/home/user/workdir \
  --net=host \
  --privileged \
  pinto0309/openvino2tensorflow:latest

$ cd workdir

$ wget https://github.com/google/mediapipe/blob/master/mediapipe/modules/hand_landmark/hand_landmark.tflite

$ tflite2tensorflow \
--model_path hand_landmark.tflite \
--flatc_path ../flatc \
--schema_path ../schema.fbs \
--output_pb

$ tflite2tensorflow \
--model_path hand_landmark.tflite \
--flatc_path ../flatc \
--schema_path ../schema.fbs \
--output_no_quant_float32_tflite

$ $INTEL_OPENVINO_DIR/deployment_tools/model_optimizer/mo_tf.py \
--saved_model_dir saved_model \
--output_dir saved_model/openvino/FP32

$ openvino2tensorflow \
--model_path saved_model/openvino/FP32/saved_model.xml \
--output_h5

• hand_landmark.tflite to .h5 converted model / .pb / saved_model / float32 tflite
  https://drive.google.com/file/d/16TTLpcZxl-llunxSJc2zQ42LSnoZSWEw/view?usp=sharing

[MaHoef]

Hello Pinto,
thanks for the hint with the old format. Even for tf2 Xilinx uses this format:
vai_c_tensorflow2 -m ./tf2_resnet50_imagenet_224_224_7.76G_1.3/quantized/quantized.h5 -a ./arch.json -n resnet50

Thanks for providing the files via google drive, really nice from you.

I tried to reproduce exactly what you did. Very strange as I still get the error:
../flatc: error: binary "hand_landmark.tflite" does not have expected file_identifier "TFL3"

I attached the full log, maybe you have an idea
tflite2tensorflow.log

[PINTO0309]

Can you successfully visualize the tflite file you downloaded with the tool at the URL below? There is a good chance that the file you downloaded is corrupt.
https://netron.app/

• hand_landmark.tflite
  

[MaHoef]

You were absolutely right, the image was corrupt. Wget is not the tool to download something from github.
It works now as expected.
I will proceed now to get the converted model compiled by the Xilinx tools which might be a challenge.
Thank you
Marco