search

PINTO0309 / openvino2tensorflow

This script converts the ONNX/OpenVINO IR model to Tensorflow's saved_model, tflite, h5, tfjs, tftrt(TensorRT), CoreML, EdgeTPU, ONNX and pb. PyTorch (NCHW) -> ONNX (NCHW) -> OpenVINO (NCHW) -> openvino2tensorflow -> Tensorflow/Keras (NHWC/NCHW) -> TFLite (NHWC/NCHW). And the conversion from .pb to saved_model and from saved_model to .pb and from .pb to .tflite and saved_model to .tflite and saved_model to onnx. Support for building environments with Docker. It is possible to directly access the host PC GUI and the camera to verify the operation. NVIDIA GPU (dGPU) support. Intel iHD GPU (iGPU) support.
MIT License
334 stars 40 forks source link

Unable to convert OpenVino Facenet-20180408-102900 model to tflite #56

Closed maurobento closed 3 years ago

maurobento commented 3 years ago

Hi, first of all thanks for your work, it's amazing!

Im using the docker image in Ubuntu 20.04 to convert the Facenet-20180408-102900 model in OpenVino format to tflite but i'm having some issues that i cannot resolve.

I added a json file to correct the errors due to reshape and transpose operations The json file looks like this:

{
    "format_version": 1,
    "layers": [
        {
            "layer_id": "728",
            "replace_mode": "direct",
            "values": [
                1,
                1,
                512,
                512
            ]
        }
    ]
}

But then, when i try to convert the model this massege comes up:

The NormalizeL2 layer is not yet implemented.

It's obvious that the NormalizeL2 operation is not implemented yet, but maybe you have a workarround to this problem.

Where can i send you the .bin and .xml files so you can try them?

Thanks

Mauro

PINTO0309 commented 3 years ago

Please provide a model if you can. I can't reproduce the problem.

maurobento commented 3 years ago

Hi, i sent you an email to this addres "rmsdh122@yahoo.co.jp" with a link to download the network because is to heavy to attach here. Let mi know if there is any problem downloading the files.

Thanks

PINTO0309 commented 3 years ago

If you're using this model, which is MIT licensed, then I see no point in you hiding the URL. Anyway, I will check the conversion. https://docs.openvinotoolkit.org/latest/omz_models_model_facenet_20180408_102900.html

PINTO0309 commented 3 years ago

Fixes: cefbef972efdc47861ab4c1cbd936bca812d49f1 https://github.com/PINTO0309/openvino2tensorflow/releases/tag/v1.16.0